{
  "id": "cognitive-psychology",
  "title": "Cognitive Psychology",
  "lang": "en",
  "source_lang": "en",
  "summary": "The scientific study of mental processes involved in perception, attention, memory, language, learning, and thinking.",
  "category": "society",
  "subcategory": "psychology",
  "type": "concept",
  "aliases": [
    "cognitive information-processing psychology"
  ],
  "generated_at": "2026-10-22",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/cognitive-psychology/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/cognitive-psychology/",
      "json": "https://aiwiki.page/zh/cognitive-psychology.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Cognitive psychology is the branch of psychology concerned with how people acquire, represent, transform, retain, and use information. Its subjects include perception, attention, memory, learning, language, reasoning, problem solving, and decision making. It investigates mental processes through observable evidence, especially patterns of accuracy, response time, and behavior under controlled conditions. Cognition includes processes that operate both within and outside conscious awareness. (dictionary.apa.org)\n\nThe field developed from earlier experimental traditions and became a distinct research movement during the mid-twentieth-century cognitive revolution, influenced by information processing, linguistics, and computing. It is a major contributor to cognitive science, alongside disciplines such as neuroscience, linguistics, and artificial intelligence. Its explanations connect experimental findings with theories of mental representation and processing, sometimes formalized as computational models. (nobaproject.com)\n\nCognitive psychology provides frameworks for understanding both the capabilities and limitations of human thought. Its findings inform education, interface design, legal investigations, and other activities that depend on how people attend, understand, remember, and choose. The field encompasses research on individual mental functions, their coordination, their development, and the conditions under which experimental findings generalize to everyday life. (psychology.stanford.edu)"
    },
    {
      "heading": "Subject Matter and Disciplinary Boundaries",
      "text": "Cognitive psychology is the branch of psychology concerned with how information is acquired, represented, transformed, stored, and used. Its subject matter includes perception, attention, memory, language, reasoning, and decision making. In this context, a representation is an internal state that carries information about something—for example, the sound of a spoken word, the appearance of an object, or a remembered instruction. Cognitive explanations specify operations on such information: selecting relevant input, relating it to existing knowledge, maintaining it during a task, or retrieving it to guide action. These descriptions concern the organization of mental activity, rather than merely cataloguing the situations in which a particular behavior occurs. (apa.org)\n\nA cognitive process is not identical to its observable outcome. Pressing a button to indicate that a symbol was previously presented is a response; encoding the symbol, retaining information about it, comparing it with a test item, and selecting an answer are proposed processes contributing to that response. The distinction makes unobservable mechanisms scientifically investigable: a theory must predict measurable differences under specified conditions. In a 1966 experiment, Saul Sternberg asked participants whether a test symbol belonged to a short memorized sequence. Mean response time increased approximately linearly with sequence length; he interpreted the slope as evidence for serial comparison at roughly 25–30 symbols per second. The explanatory claim concerned an internal operation, whereas the evidence consisted of response-time patterns. Such inference goes beyond introspection by making claims answerable to controlled observations rather than to subjective description alone. (dictionary.apa.org)\n\nCognitive science is the broader interdisciplinary enterprise within which cognitive psychology participates. It brings psychological research into contact with linguistics, artificial intelligence, neuroscience, philosophy, anthropology, and education. Its questions can concern the structure of a language, the computational requirements of intelligent action, or the relationship between mental representations and knowledge, rather than human experimental performance alone. Cognitive psychology contributes especially by developing and testing accounts of how people actually perform cognitive tasks. The distinction is therefore one of disciplinary scope, not a division between unrelated subjects: psychological findings constrain interdisciplinary theories, while ideas from computation and language research help formulate psychological hypotheses. (cognitivesciencesociety.org)\n\nCognitive neuroscience places greater emphasis on how cognitive operations are implemented in the brain. A functional question asks what information must be maintained or transformed to accomplish a task; an implementation question asks how neural activity realizes those operations. Cognitive neuroscience therefore relates behavioral performance to brain measurements and tests models against both kinds of evidence. These are complementary explanatory aims, not mutually exclusive research territories: the same computational model may generate predictions about responses and neural activity. Identifying an operation at the psychological level and explaining its biological realization address different aspects of the same capacity. (arxiv.org)\n\nBasic cognitive research is also distinct from clinical psychology, whose remit includes the prevention, assessment, diagnosis, and treatment of mental, emotional, and behavioral disorders. The distinction concerns research aims and professional functions, not whether either field uses experiments. Cognitive-behavioral therapy is a psychological treatment approach involving changes in patterns of thinking and behavior; it is neither another name for cognitive psychology nor its defining activity. For example, investigating how information is retained seeks an explanation of a cognitive capacity, whereas helping a patient identify and reevaluate unhelpful interpretations serves a therapeutic purpose. Clinical psychology encompasses multiple theoretical approaches, and cognitive research extends well beyond questions of psychological disorder or treatment. (apa.org)"
    },
    {
      "heading": "Historical Foundations and the Cognitive Revolution",
      "text": "Nineteenth-century experimental research established ways to measure mental activity before cognitive psychology became a named field. Ernst Heinrich Weber investigated sensory discrimination; Gustav Theodor Fechner’s Elemente der Psychophysik (1860) developed psychophysics as a quantitative study of relationships between physical stimulation and sensation. Wilhelm Wundt’s laboratory, established at Leipzig in 1879, provided an institutional base for controlled psychological experiments and international research training. These developments made perception and conscious experience subjects of systematic measurement rather than philosophical discussion alone. (lw.uni-leipzig.de) Hermann Ebbinghaus extended experimentation to memory in Über das Gedächtnis (1885), using himself as a participant and learning lists of nonsense syllables. His “savings” measure compared the work required for initial learning with that required for relearning after a delay, detecting retention even when material could no longer be reproduced immediately. (psychclassics.yorku.ca)\n\nBehaviorism redirected psychological explanation toward observable behavior and the conditions governing it. John B. Watson’s 1913 manifesto rejected introspection as an essential method and identified prediction and control of behavior as psychology’s goals. Learning research subsequently emphasized relationships among environmental events, responses, and consequences. Nevertheless, behaviorist traditions were not uniformly opposed to inferred internal organization: Edward C. Tolman’s 1948 account of “cognitive maps” interpreted rats’ maze performance in terms of acquired spatial relationships rather than only reinforced sequences of movements. His work supplied an important bridge between experimental learning research and later representational explanations. (psychclassics.yorku.ca)\n\nGestalt psychology maintained a contrasting emphasis on organized experience. Max Wertheimer, Wolfgang Köhler, and Kurt Koffka argued that perception and problem solving could not be adequately explained by treating their elements independently. Wertheimer’s 1923 investigations of perceptual grouping showed, for example, how proximity and similarity determine which marks appear to belong together. The relevant explanatory unit was the structured configuration, not merely an inventory of sensations. Gestalt research thus preserved questions about internal organization during the period when behaviorism was especially influential in American psychology. (nobaproject.com)\n\nThe mid-twentieth-century cognitive revolution brought several research programs into contact rather than replacing behaviorism through a single discovery. Information theory supplied quantitative concepts of information and channel capacity, while programmable computers demonstrated how explicit operations on representations could produce complex performances. George A. Miller’s 1956 paper, “The Magical Number Seven, Plus or Minus Two,” connected experimental findings about capacity limits with information processing and the recoding of items into chunks. At an MIT information-theory symposium in September 1956, work by Miller, Noam Chomsky, Allen Newell, and Herbert A. Simon brought experimental psychology, formal language research, and computer-based reasoning into a shared discussion. Miller later identified this meeting as a formative convergence, not evidence that all these approaches had a single origin. (cs.princeton.edu)\n\nJerome Bruner, Jacqueline Goodnow, and George Austin’s A Study of Thinking (1956) investigated how people formed concepts through active strategies rather than passive accumulation of associations. In language research, Chomsky emphasized structured grammatical knowledge capable of supporting novel utterances, challenging explanations based primarily on learned verbal responses. Newell and Simon’s computer simulations made reasoning procedures explicit enough to execute and examine. These programs differed in subject matter, but each treated internal structure or operations as legitimate components of experimentally constrained explanations. (psychology.fas.harvard.edu)\n\nInstitutional consolidation accompanied these theoretical changes. Bruner and Miller founded Harvard’s Center for Cognitive Studies in 1960, creating an interdisciplinary setting for research on cognition. (psychology.fas.harvard.edu) Ulric Neisser’s Cognitive Psychology (1967) synthesized research on perception, attention, memory, and related information-processing activities under a recognizable disciplinary title. Its opening chapter drew together experimental findings with material from programming, linguistics, and other psychological specialties; the book consolidated an existing movement rather than founding it independently. (stanford.edu) The specialist journal Cognitive Psychology, which began publication in 1970, provided a further institutional outlet for this research. (ncbi.nlm.nih.gov)"
    },
    {
      "heading": "Experimental Methods and Behavioral Measurement",
      "text": "Cognitive experiments translate theoretical questions into operational definitions: explicit procedures for manipulating conditions and measuring performance. An independent variable is manipulated by the researcher; a dependent variable is the recorded outcome. In J. Ridley Stroop’s 1935 experiments, participants named ink colors while ignoring conflicting color words, or named comparable colors presented as solid squares. Stimulus format was the manipulation, and completion time and errors were behavioral outcomes. Naming 100 colors took an additional 47 seconds—an increase of 74.3%—when the colors appeared as conflicting words rather than squares. The time difference operationalized interference; it was not a direct measurement of a separately observable mental process. (nobaproject.com)\n\nExperimental design determines which alternative explanations a comparison can exclude. In a between-participant design, different people receive different conditions; random assignment makes allocation independent of pre-existing characteristics in expectation, although it does not guarantee identical groups in a particular sample. In a within-participant design, each person completes multiple conditions, controlling for stable differences in baseline performance. Repeated testing, however, introduces practice, fatigue, and carryover effects. Randomizing trial order or counterbalancing condition order distributes these influences rather than allowing one condition always to occur last. Standardized instructions and comparable materials further reduce confounding. A correlation between two measured abilities does not, by itself, establish that either causes the other; controlled manipulation supports causal inference by separating the manipulated factor from competing influences. (nobaproject.com)\n\nResponse time measures the interval between a specified event, such as stimulus onset, and a response; accuracy records correct judgments or particular error types. Their value often lies in patterns across conditions rather than an isolated score. In Saul Sternberg’s 1966 memory experiment, participants judged whether a test symbol belonged to a short memorized sequence. Mean response time increased approximately linearly with sequence length, supporting his proposal of comparisons occurring at about 25–30 symbols per second. Such measurements constrain explanations by requiring them to account for how performance changes with task demands, not merely whether participants succeed. (pubmed.ncbi.nlm.nih.gov)\n\nMemory tasks distinguish recall, which requires producing studied information, from recognition, which requires identifying previously encountered items among alternatives. These procedures provide different retrieval cues and therefore need not yield equivalent estimates of memory. Eye tracking adds temporally localized measures, including fixation duration, fixation probability, and gaze duration on a word before the eyes move onward. Erik Reichle and colleagues’ 1998 E-Z Reader model used these distinct outcomes to test relationships between word processing and eye-movement control. Combining response times, errors, memory responses, and gaze measures constrains interpretation more effectively than treating any single indicator as a transparent measure of cognition. (nobaproject.com)\n\nPsychophysical tasks systematically vary physical stimulus properties to estimate relationships between stimulation and judgment. A two-alternative forced-choice task, for example, requires choosing between two possible responses, giving chance accuracy of 50%. A psychometric function relates stimulus intensity to response probability; its parameters can distinguish threshold, slope, guessing, and lapses. Threshold estimates therefore depend on a specified task and performance criterion rather than identifying an absolute boundary below which perception invariably disappears. Interpretation also requires accounting for the speed–accuracy trade-off: faster responses may reflect reduced caution rather than improved processing. Instructions and deadlines can change both response time and error rate, making their joint examination essential. (pmc.ncbi.nlm.nih.gov)\n\nMeasurement reliability concerns consistency, including whether repeated testing preserves individuals’ relative scores. It is distinct from the size or dependability of an average experimental effect. Craig Hedge, Georgina Powell, and Petroc Sumner demonstrated this distinction in their 2018 study of seven widely used cognitive tasks: robust group-level effects could coexist with weak reliability for individual differences. Tasks that elicit similar condition effects across participants can leave little stable between-person variation to measure. Consequently, an average interference effect does not establish that the corresponding difference score reliably ranks individuals, nor that every participant exhibits the group’s average performance pattern. (pubmed.ncbi.nlm.nih.gov)"
    },
    {
      "heading": "Explanatory Frameworks and Computational Models",
      "text": "Information-processing models explain cognition as transformations of representations: a stimulus is encoded, relevant knowledge is retrieved, a response is selected, and an action is executed. Serial operations occur successively; parallel operations overlap. Limited capacity can be modeled as a restricted buffer, competition among representations, or a bottleneck allowing only one operation at a time. These are distinct mechanisms, not interchangeable descriptions of mental effort. A system can therefore process sensory information in parallel while selecting actions serially, as in the ACT-R architecture. Such proposals specify where interference should arise rather than merely asserting that cognition has limited resources. (cs.utexas.edu)\n\nStage models become testable when their operations generate numerical predictions. In Saul Sternberg’s 1966 memory-scanning experiment, participants judged whether a probe belonged to a short memorized sequence. Mean response time increased approximately linearly with sequence length; Sternberg interpreted the slope as a serial comparison rate of 25–30 symbols per second. This interpretation distinguished the time spent comparing items from other components of the response. Information flow need not, however, proceed only upward from sensory input. Bottom-up influences originate in stimulus information, whereas top-down influences reflect higher-level representations. James L. McClelland and David E. Rumelhart’s 1981 interactive-activation model implemented both: visual features activated compatible letters, letters activated words, and words fed activation back to their constituent letters. Excitation and competition across simultaneously active levels accounted for contextual advantages in letter identification. (pubmed.ncbi.nlm.nih.gov)\n\nSymbolic accounts represent knowledge using identifiable structures and manipulate them through explicit rules. A production rule, for example, specifies an action when particular conditions match the current goal and available information. Connectionist accounts instead describe interacting processing units whose activation depends on weighted connections. In distributed representations, a concept corresponds to a pattern across multiple units rather than one dedicated symbol. Learning changes connection strengths, thereby changing the network’s responses. Rumelhart, Geoffrey Hinton, and Ronald Williams’s 1986 neural-network study demonstrated how back-propagated error could adjust weights to reduce differences between actual and target outputs; hidden units acquired features useful for the task. This contrasts with directly specifying the relevant features and transformation rules in advance. (cs.utexas.edu)\n\nProbabilistic inference explains cognition in terms of uncertainty over possible interpretations or hypotheses. By Bayes’ theorem, \\(P(h\\mid d)\\propto P(d\\mid h)P(h)\\): the posterior probability of hypothesis \\(h\\), given data \\(d\\), depends on its prior probability and the likelihood of those data under it. The framework makes assumptions about prior knowledge and hypothesis spaces explicit. Crucially, a normative model specifies an optimal solution under stated assumptions; a descriptive model explains what people actually do. Agreement with an ideal Bayesian solution does not by itself identify the mental algorithm. Thomas Griffiths and colleagues’ 2010 treatment distinguished ideal inferential solutions from algorithms that approximate them and neural mechanisms that might implement those algorithms. (digital.library.adelaide.edu.au)\n\nACT-R, developed by John R. Anderson and collaborators, illustrates a cognitive architecture integrating several functions within shared constraints. Its 2004 formulation coordinated perceptual, motor, goal, and declarative-memory modules through limited-capacity buffers and production rules. Modules could operate concurrently, but each buffer held one knowledge unit, or “chunk,” and only one production was selected per cycle, modeled as approximately 50 milliseconds. ACT-R combined symbolic chunks and rules with numerical activation and utility mechanisms governing retrieval and selection. In a simulated radar-monitoring task, the architecture generated predictions about component response times, practice-related improvement, and task-relevant fixation patterns. Remembering menu locations reduced visual search, while production compilation shortened action sequences. These process-specific predictions distinguish an explanatory model from a program that merely completes the task: successful output alone does not establish that the intervening operations resemble human cognition. (cs.utexas.edu)"
    },
    {
      "heading": "Perception and Object Recognition",
      "text": "Sensory registration converts physical stimulation into neural signals: retinal photoreceptors respond to light, while cochlear hair cells transduce mechanical vibrations associated with sound. Perception involves organizing and interpreting those signals as objects, surfaces, and events. Detecting differences in light across an image does not itself identify a cup; registering successive tones does not determine whether they belong to one melody or separate sound sources. In auditory-streaming experiments, alternating high and low tones can be heard as one sequence or as two streams, depending on frequency separation and presentation rate. The organization of experience therefore depends on relationships among signals, not merely their detection. (nobaproject.com)\n\nVisual grouping similarly determines which elements belong together. Proximity favors grouping nearby elements, similarity groups elements sharing properties such as color, and good continuation favors smooth contours over abrupt changes in direction. Figure–ground organization assigns a boundary to a shaped foreground region rather than to the background extending behind it; computational models treat this “border ownership” as a problem requiring interactions between local edges and wider scene information. Depth perception combines binocular disparity—the difference between the eyes’ images—with information from occlusion and motion parallax. These processes organize a two-dimensional retinal projection into an interpretation of surfaces arranged in three-dimensional space. (cs.utexas.edu)\n\nPerceptual constancy concerns relative stability despite changing sensory input. A person moving away produces a smaller retinal image without necessarily appearing physically smaller. In their 1941 experiments, Alfred Holway and Edwin Boring manipulated available distance information: size judgments became more dependent on retinal image size when distance cues were restricted. Context also affects identification. Stephen Palmer’s 1975 experiment presented scenes before briefly displayed object drawings; identification was most accurate with an appropriate scene and least accurate with an inappropriate one. A kitchen scene, for example, supplied an appropriate context for bread but not a mailbox. Such results establish effects on recognition performance, without by themselves specifying whether context changes early visual processing, later interpretation, or response selection. (wexler.free.fr)\n\nObject recognition must accommodate changes in viewpoint, lighting, scale, and partial visibility. Irving Biederman’s 1987 recognition-by-components theory proposed that objects are represented through arrangements of simple volumetric components, or geons, including cylinders, blocks, and wedges. Its proposed vocabulary of approximately 36 components depends on edge properties such as curvature, parallelism, and symmetry that often survive viewpoint changes. Recognition can therefore remain possible when visible contours preserve diagnostic components and their spatial relationships, rather than matching a complete retinal image. This predicts particular difficulty when an unusual view hides the components needed to distinguish an object. (citeseerx.ist.psu.edu)\n\nFace recognition illustrates the importance of whole-pattern organization when distinguishing individual members of a visually similar category. In three experiments published in 1993, James Tanaka and Martha Farah found that participants identified a facial feature more accurately within a whole upright face than in isolation, even when the alternatives differed only in that feature. Scrambled faces, inverted faces, and houses did not show the same whole-object advantage. The result supports a holistic contribution to upright-face recognition: the perceptual context supplied by the complete face affects identification of its parts. It does not imply that individual features are irrelevant. (pubmed.ncbi.nlm.nih.gov)\n\nMultisensory integration shows that interpretation can also combine different senses. Harry McGurk and John MacDonald reported in 1976 that an auditory /ba/ paired with visible articulation of /ga/ could produce a perceived /da/, although the soundtrack alone was identified correctly. This illusion experimentally separates acoustic input from perceived speech. In 2002, Marc Ernst and Martin Banks measured visual and haptic judgments of height and found that combined judgments closely followed a model weighting each estimate according to its reliability: vision contributed less when its uncertainty increased. Such findings distinguish integration from a fixed hierarchy in which vision always dominates, and make perceptual discrepancies evidence about identifiable processing mechanisms. (nature.com)"
    },
    {
      "heading": "Attention and Selection",
      "text": "Selective attention prioritizes some information over competing inputs: experiments ask which properties guide selection and how far ignored material is processed. Divided attention concerns coordinating concurrent tasks, typically assessed by comparing response speed and accuracy under single- and dual-task conditions. Sustained attention concerns maintaining task engagement over time, especially when targets are rare; vigilance experiments measure changes in detection, false alarms, and response times during prolonged monitoring. These distinctions concern task demands rather than three wholly separate faculties: monitoring for rare signals can also require selecting relevant inputs and coordinating responses. (nobaproject.com)\n\nThe cocktail-party problem concerns following one speaker among competing voices. In dichotic-listening experiments, different messages are delivered simultaneously to the two ears, and participants “shadow”—repeat aloud—one message. They typically recall little of the ignored message’s meaning while retaining some physical information, such as voice characteristics. Donald Broadbent’s 1958 early-selection model proposed a filter based on physical properties before capacity-limited identification. Anne Treisman’s attenuation account instead weakened unattended inputs, allowing sufficiently important or readily activated words to influence processing. Late-selection accounts, associated with J. Anthony Deutsch and Diana Deutsch in 1963, placed selection after semantic analysis, at access to response or report. Occasional detection of one’s name in an ignored stream challenges an absolute filter, but does not establish that every unattended message receives complete semantic analysis. (nobaproject.com)\n\nSpatial-cueing experiments distinguish directing attention from moving the eyes. In Michael Posner’s 1980 work, participants maintained fixation while cues indicated where a target would probably appear. One procedure used arrows that were valid on 80% of cued trials and invalid on 20%. Faster detection at expected locations demonstrated covert orienting; misleading cues required attention to be redirected. Predictive cues examine goal-directed selection, whereas conspicuous, task-irrelevant events test stimulus-driven attentional capture. These influences interact: Charles Folk, Roger Remington, and James Johnston reported in 1992 that abrupt-onset distractor cues disrupted performance when targets were defined by onset, but not when targets were defined by color. Capture therefore depended on the properties participants were prepared to select. (www2.psychology.uiowa.edu)\n\nVisual search measures how efficiently a target is found among distractors as display size changes. In Anne Treisman and Garry Gelade’s 1980 experiments, detecting a distinctive single feature produced relatively little increase in response time with additional items, whereas finding a conjunction of features produced larger increases. Their feature-integration theory proposed parallel registration of elementary features followed by focused attention that binds features into objects. For example, a red target among green distractors differs by color alone; a red vertical target among red horizontal and green vertical items requires combining color and orientation. However, later experiments demonstrated parallel processing in both feature and conjunction searches, showing that search slopes alone do not uniquely establish item-by-item scanning. (cseweb.ucsd.edu)\n\nInattentional blindness illustrates failure to notice an unexpected visible event while attention is engaged elsewhere. In Daniel Simons and Christopher Chabris’s 1999 study, observers counted basketball passes while an unexpected person crossed the scene. Across 192 analyzed observers and 16 conditions—including gorilla-costume and umbrella-carrying events—46% failed to report the event; this percentage was not a universal “gorilla miss rate.” Detection varied with display format, counting difficulty, and similarity to the attended team. The experiment demonstrated limits on reportable awareness under specified attentional demands, rather than an absence of all processing of unreported objects. (chabris.com)\n\nDual-task interference can arise even when both stimuli are expected. In the psychological refractory-period paradigm, two stimuli require separate speeded responses; shortening their separation typically delays the second response. Harold Pashler’s 1994 synthesis identified evidence for a central bottleneck involving response selection, while distinguishing additional perceptual, preparation, and timing limitations. Such results do not imply that all processing is serial: perceptual operations and response execution may overlap while particular central operations remain constrained. Laboratory interference consequently supports claims about specified task combinations, not a blanket assertion that every form of everyday multitasking is impossible or equally costly. (citeseerx.ist.psu.edu)"
    },
    {
      "heading": "Short-Term Retention and Working Memory",
      "text": "Short-term memory refers to the brief retention of information, whereas working memory emphasizes keeping information available for use during ongoing thought and action. Remembering a digit sequence long enough to repeat it primarily tests retention; retaining intermediate results while calculating a total also requires updating and coordination. The distinction is functional rather than simply temporal: working memory is not defined by a longer retention interval, and its contents need not undergo manipulation continuously. Terminology varies across theories, with some treating short-term storage as a component of working memory and others emphasizing the processes that make representations temporarily accessible. (pmc.ncbi.nlm.nih.gov)\n\nAlan Baddeley and Graham Hitch’s 1974 multicomponent account distinguished an attentional controller, the central executive, from systems specialized for verbal–acoustic and visuospatial information. In the developed model, the phonological loop combines temporary phonological storage with articulatory rehearsal, while the visuospatial sketchpad maintains visual patterns and spatial relationships. These components allow partially independent maintenance, although concurrent activities can still compete for executive control. In 2000, Baddeley added the episodic buffer, a limited-capacity system proposed to bind information from the subsidiary systems and long-term memory into integrated representations—for example, combining words with their visual and semantic context. These are functional components, not claims that each occupies one isolated brain location. (doi.org)\n\nNelson Cowan’s embedded-processes account instead emphasizes temporarily activated long-term representations and a more restricted focus of attention within that activated information. Executive processes regulate which representations receive attention and how they are used. This approach does not require every temporarily available representation to be copied into a separate short-term store. Familiar word meanings, for example, can become activated while only a few currently relevant units occupy attentional focus. The disagreement with multicomponent accounts therefore concerns the organization of temporary accessibility and the source of capacity limits, not whether prior knowledge and attention matter. Specialized stores and activated-representation accounts also need not be mutually exclusive in every detail. (pmc.ncbi.nlm.nih.gov)\n\nCapacity is measured through tasks with different demands. Simple span tasks require immediate reproduction of progressively longer sequences, usually in presentation order. Complex span tasks interleave retention with processing: reading span combines sentence processing with recall of designated words, while operation span alternates memory items with arithmetic. Chunking changes the effective units being retained by grouping elements into familiar structures; consequently, remembering four familiar abbreviations is not equivalent to retaining the same number of unrelated letters. George A. Miller’s 1956 discussion of approximately seven chunks did not establish a universal seven-item storage limit. Cowan’s 2001 analysis argued for approximately four chunks under conditions restricting rehearsal and recoding. Such estimates depend on the materials, prior knowledge, presentation conditions, scoring, and opportunities for strategic grouping. (pmc.ncbi.nlm.nih.gov)\n\nRehearsal and interference can be investigated experimentally rather than inferred solely from span scores. Repeating an irrelevant syllable during presentation—articulatory suppression—occupies speech-production processes and disrupts verbal maintenance. Phonologically similar items are generally harder to reproduce in order than dissimilar items; suppression can remove this similarity effect for visually presented material, although it commonly persists with auditory presentation. This interaction supports a role for speech-based recoding without showing that all short-term retention is verbal. Forgetting can also reflect competition among items or their ordering information. Distinguishing interference from time-based decay requires controlling what happens during a delay, because longer delays often introduce additional distracting activity. (pmc.ncbi.nlm.nih.gov)\n\nWorking-memory performance is associated with language comprehension and reasoning, but these associations do not identify a single causal mechanism. Meredyth Daneman and Patricia Carpenter’s 1980 reading-span experiments linked combined processing-and-storage performance to comprehension, including retrieving facts and resolving pronoun references. Subsequent research has examined attentional control and mind wandering as contributors to this relationship. Maintaining task goals, resisting distraction, retrieving relevant knowledge, and efficiently processing sentences can all affect measured performance. A span score therefore reflects the interaction of retention with task-specific operations and strategies, rather than a direct reading of one context-independent mental container. (eric.ed.gov)"
    },
    {
      "heading": "Long-Term Memory and Forgetting",
      "text": "Long-term memory includes distinguishable forms of retained information. Episodic memory concerns events situated in a particular time and place, such as a conversation at yesterday’s lunch; semantic memory concerns facts and meanings, such as knowing what a restaurant is, without necessarily remembering when that knowledge was acquired. Procedural memory supports skills expressed through performance, such as mirror drawing. Episodic and semantic memory are commonly grouped as declarative memory, whereas procedural learning belongs to a broader, heterogeneous category of nondeclarative memory. Explicit and implicit measures distinguish how retention is tested: explicit tests request deliberate remembering, while implicit tests reveal effects of previous experience without requiring recollection—for example, improved identification of previously encountered words. Implicit memory is therefore not simply another name for procedural memory. (nobaproject.com)\n\nRemembering can be analyzed through encoding, storage, and retrieval: establishing a representation, retaining information over time, and accessing it subsequently. These are interdependent processes rather than isolated compartments. Elaborative encoding relates new information to existing knowledge—for example, connecting an unfamiliar occupation to its characteristic activities—thereby establishing potentially useful retrieval routes. A retrieval cue is effective partly because of its relationship to the original encoding. Endel Tulving and Donald Thomson’s 1973 encoding-specificity principle formalized this dependence: a cue’s usefulness cannot be predicted solely from its ordinary association with the target. Their experiments demonstrated that words not recognized on one test could nevertheless be recalled when an appropriate studied cue was supplied. (nobaproject.com)\n\nRecall requires producing previously encountered information, either with minimal prompting in free recall or with assistance in cued recall. Recognition presents candidate items and requires distinguishing previously encountered targets from new alternatives. Recognition often benefits from the target’s presence as a cue, but the recognition-failure findings show why it cannot invariably be ranked as easier than recall. Forgetting likewise does not necessarily establish that stored information has disappeared: successful retrieval after a reminder indicates that an earlier failure reflected limited accessibility. Proposed causes include inadequate initial encoding, deterioration of memory traces, disrupted consolidation, and competition among memories. Proactive interference occurs when earlier learning obstructs newer learning or retrieval; retroactive interference occurs when subsequent learning obstructs retention of earlier material. Learning a replacement password, for example, can create competition with its predecessor. Elapsed time alone does not distinguish decay from intervening activity or changing retrieval conditions. (nobaproject.com)\n\nRemembering is reconstructive: retrieved details are interpreted alongside semantic associations, expectations, and judgments about their origin. Marcia Johnson, Shahin Hashtroudi, and D. Stephen Lindsay’s 1993 source-monitoring framework explains how people attribute remembered information to perception, imagination, or another person’s account, using features such as perceptual detail and contextual information. Source errors occur when content is retained but its origin is misidentified. In Henry Roediger and Kathleen McDermott’s 1995 experiments, participants studied twelve-word lists containing associates such as bed, rest, and awake, but not sleep. In the first experiment, the omitted associate was nevertheless produced on 40% of immediate free-recall opportunities and was subsequently recognized with high confidence. This laboratory demonstration of false memory shows that associative coherence can support confident errors; it does not establish that every confident recollection is inaccurate. (memlab.yale.edu)\n\nOrdinary omissions and source errors differ from clinical amnesia, which can involve severe impairment in acquiring new declarative memories—anterograde amnesia—or loss of information acquired before onset—retrograde amnesia. Neither necessarily entails the loss of every memory function. Following bilateral medial temporal surgery in 1953, the patient H.M. showed profound difficulty establishing new declarative memories. Yet Brenda Milner’s testing demonstrated improvement in mirror drawing across three days despite his lack of recollection of previous practice. Such dissociations provided evidence that successful skill retention and conscious remembering can depend on separable memory processes. (pmc.ncbi.nlm.nih.gov)"
    },
    {
      "heading": "Learning, Practice, and Knowledge Acquisition",
      "text": "Learning denotes relatively enduring changes in knowledge or skill attributable to experience, rather than every improvement observed during training. Performance can improve temporarily because material remains immediately accessible or assistance is available, without equivalent gains on a delayed, unassisted test. Conversely, learning may occur without an immediately visible improvement. Researchers therefore distinguish acquisition performance from later retention and transfer: a practice condition that produces fewer errors initially need not produce more durable knowledge. This distinction is central to experimental comparisons of repetition, retrieval, and practice schedules. (psychologicalscience.org)\n\nIntentional learning involves an explicit goal of acquiring information; incidental learning occurs while pursuing another activity, such as judging words without expecting a subsequent memory test. This distinction concerns intention, whereas implicit learning concerns acquisition expressed without reportable awareness of the acquired regularity; incidental learning need not be implicit. In Mary Jo Nissen and Peter Bullemer’s 1987 serial reaction-time experiments, participants responded to stimulus locations following a repeating ten-trial sequence. Performance measures revealed sequence learning, including in participants who could not report the pattern, while a concurrent task substantially reduced acquisition. Perceptual learning instead concerns experience-dependent improvements in discrimination. Avi Karni and Dov Sagi’s 1991 texture-discrimination study found durable improvement that was specific to retinal location and background orientation, with little transfer between eyes. These forms of learning are distinguished by their experimental measures, not simply by whether participants received instruction. (nobaproject.com)\n\nAssociative learning involves acquiring relationships between events or between actions and outcomes, but temporal co-occurrence alone does not determine what is acquired. In blocking experiments, prior training that establishes cue A as a predictor of an outcome reduces subsequent learning about cue B when the compound AB predicts the same outcome. Robert Rescorla and Allan Wagner’s 1972 model explains this through prediction error: associative change depends on the discrepancy between the observed outcome and the outcome predicted jointly by the available cues. An already expected outcome produces little additional learning. Prior knowledge also shapes acquisition by making new information interpretable. In John Bransford and Marcia Johnson’s 1972 experiments, supplying an appropriate context before an otherwise obscure passage increased comprehension and recall substantially more than supplying that context afterward. (pmc.ncbi.nlm.nih.gov)\n\nPractice experiments manipulate not only repetition count but also spacing, intervening activities, and the delay before assessment. Distributed practice separates learning episodes in time, whereas massed practice places them close together. Nicholas Cepeda and colleagues’ 2008 study taught facts to more than 1,350 participants, varied the interval before review up to approximately 3.5 months, and tested retention after delays extending to one year. At a given test delay, increasing the study interval first improved and then reduced retention; the best-performing interval increased with the intended retention period. Spacing therefore does not imply that indefinitely longer gaps are better, or that one schedule is optimal for every assessment delay. (laplab.ucsd.edu)\n\nRetrieval practice requires producing previously encountered information rather than merely encountering it again. Henry Roediger and Jeffrey Karpicke’s 2006 prose-learning experiments found that repeated study favored recall after five minutes, whereas retrieval practice favored retention after two days or one week, even without feedback during practice. Testing thus changes subsequent learning rather than only measuring it. Feedback is a separate variable: it can supply correct answers after failures and prevent incorrect responses from persisting. Andrew Butler and Roediger’s 2008 experiments showed that multiple-choice testing could also promote later production of misleading answer alternatives; corrective feedback reduced this cost while increasing correct retention. Benefits consequently depend on test format, feedback, and assessment timing, not merely on the presence of a quiz. (pubmed.ncbi.nlm.nih.gov)\n\nTransfer of learning concerns effects on an unpractised task or context. Near transfer involves closely related materials or procedures; farther transfer requires generalization across more substantial differences in content, setting, or required operations. Mary Gick and Keith Holyoak’s 1980 experiments demonstrated both the possibility and the limits of cross-domain transfer: participants who read a military story about converging forces could use its structure to solve a radiation problem requiring converging low-intensity rays, particularly when prompted to use the story. Relevant knowledge being available was therefore insufficient to ensure its spontaneous application; recognizing the correspondence and adapting the earlier solution were additional requirements. (sciencedirect.com)"
    },
    {
      "heading": "Concepts, Categories, and Semantic Knowledge",
      "text": "A category is a set of entities treated as equivalent for some purpose; a concept is a mental representation used to identify and reason about that set. Concepts allow encounters with particular objects to support expectations about unfamiliar ones. Recognizing an unfamiliar object as a chair, for example, supports an inference about sitting on it without requiring previous experience of that particular chair. Such inductive reasoning extends beyond observed features to expected functions and properties. Within semantic knowledge, the central question here concerns how meanings and categories are organized, rather than how individual experiences are encoded or retrieved. (nobaproject.com)\n\nRule-based accounts characterize membership through defining conditions: a necessary feature must be present, while a sufficient combination guarantees membership. An experimental category might contain all figures that are both red and triangular. Many everyday categories, however, resist equally satisfactory definitions. Prototype accounts instead represent a category through a summary of its characteristic features, weighted by their prevalence or diagnostic value. A new object is classified by its similarity to that summary; the prototype need not be an actual previously encountered object. This approach distinguishes knowing what birds generally look like from remembering a particular robin. (nobaproject.com)\n\nExemplar accounts retain representations of individual category members rather than requiring a single summary. In Robert Nosofsky’s 1986 generalized context model, classification depends on an object’s similarity to stored exemplars, with attention changing the importance of different stimulus dimensions. Similarities are combined across exemplars belonging to competing categories. Consequently, classification can reflect both frequently encountered examples and distinctive exceptions. Nosofsky tested the approach using identification and categorization judgments from two participants viewing the same multidimensional perceptual stimuli. Prototype and exemplar accounts can both produce typicality effects, so observing that some members are classified more readily than others does not, by itself, establish which representation is used. (pubmed.ncbi.nlm.nih.gov)\n\nTypicality concerns how representative an example seems, not simply whether it belongs. Eleanor Rosch and Carolyn Mervis’s 1975 studies connected typicality with family resemblance: good examples tend to share many attributes with other members while sharing fewer with contrasting categories. Their research included six experiments using natural and artificial categories. Overlapping similarities can therefore organize a category without one perceptible feature being shared by every member. This internal gradient must be distinguished from uncertainty at a category’s boundary: an atypical bird can still be unambiguously a bird, whereas a borderline item raises disagreement about membership itself. (worthylab.org)\n\nCategories also form hierarchies. In furniture → chair → desk chair, chair commonly functions as the basic level, between a broad superordinate and a narrower subordinate category. In 1976, Rosch and colleagues reported four experiments examining shared attributes, movements, and shapes, followed by eight investigating implications for perception, naming, and development. Basic-level categories were relatively inclusive groupings whose members nevertheless shared recognizable shapes and characteristic ways of interacting with them. “Furniture” groups objects with substantially different forms and uses; “desk chair” adds specificity within the already informative category “chair.” This advantage concerns the relationship between category structure and cognitive tasks, not merely a category’s position in a verbal hierarchy. (sciencedirect.com)\n\nKnowledge-based accounts emphasize why features belong together. Gregory Murphy and Douglas Medin argued in 1985 that background explanations help determine category coherence, rather than similarity alone supplying an adequate account. In three experiments published in 1994, Murphy and Paul Allopenna compared artificial concepts with meaningless features, meaningful but unrelated features, and features interconnected by prior knowledge. Integrated concepts were learned better; merely making features meaningful, or drawing them from the same domain without connecting them, did not provide the same advantage. Prior knowledge thus influences learning by relating attributes into an interpretable structure, not simply by making each attribute familiar. (pubmed.ncbi.nlm.nih.gov)"
    },
    {
      "heading": "Language Processing and Reading",
      "text": "Speech perception requires listeners to identify words in a continuous acoustic signal whose sound patterns vary with neighboring sounds. Lexical access—the activation and selection of stored word information—therefore involves resolving competing candidates as speech unfolds, rather than simply matching isolated sounds to dictionary entries. James McClelland and Jeffrey Elman’s 1986 TRACE model formalized this process through interacting feature, phoneme, and word units, with excitation between compatible representations and inhibition between competitors. It provides a connectionist account of how acoustic evidence and lexical knowledge jointly influence recognition. Speech production coordinates processing in the opposite direction: an intended message must be expressed through words, grammatical organization, sound sequences, and articulation. Willem Levelt, Ardi Roelofs, and Antje Meyer’s 1999 model distinguished conceptual preparation, lexical selection, morphological and phonological encoding, phonetic encoding, and articulation, using naming latencies and speech errors to constrain the proposed stages. (web.stanford.edu)\n\nSentence comprehension involves assigning grammatical relationships through syntactic parsing while constructing an interpretation through semantic processing. Temporary ambiguity makes these operations experimentally visible. In Michael Tanenhaus and colleagues’ 1995 experiments, instructions such as “Put the apple on the towel in the box” could initially encourage interpreting on the towel as a destination rather than as a description identifying the apple. Eye movements showed that the arrangement of visible objects affected this interpretation: when two apples made the identifying description useful, listeners were less likely to pursue the misleading destination analysis. Context could thus affect early parsing, not merely repair an interpretation after a sentence ended. Semantic interpretation also has a measurable time course. In 1980, Marta Kutas and Steven Hillyard found that semantically inappropriate sentence endings elicited a negative electrical response, termed the N400, distinct from responses to physically unusual print. (shotam.github.io)\n\nVisual word recognition connects written forms with pronunciation and meaning. Max Coltheart and colleagues’ 2001 dual-route cascaded model distinguished a lexical route accessing familiar word representations from a nonlexical route converting letter patterns into sounds. This distinction explains how readers can pronounce both familiar irregular words, such as pint, and unfamiliar pronounceable strings: applying common spelling–sound correspondences alone would mispronounce some exceptions, whereas stored whole-word knowledge alone would not accommodate novel strings. The model generated predictions for both reading aloud and lexical decision, in which participants classify letter strings as words or nonwords. Its routes are theoretical processing mechanisms, rather than a claim that every act of reading follows one exclusive pathway. (pubmed.ncbi.nlm.nih.gov)\n\nReading comprehension extends beyond recognizing successive words. Readers must connect propositions, resolve references, and relate textual information to relevant knowledge. Walter Kintsch’s 1988 construction–integration model distinguished an initial construction process, which activates word meanings, propositions, and associated inferences, from an integration process that strengthens mutually coherent information within the resulting network. The account distinguishes a representation of what the text explicitly states from a situation model incorporating the events or circumstances described. Understanding a passage therefore depends not only on decoding its sentences but also on establishing relationships across them and integrating their content with knowledge. (condor.depaul.edu)\n\nResearch on language acquisition connects these mechanisms with developmental psychology and linguistics. In 1996, Jenny Saffran, Richard Aslin, and Elissa Newport showed that eight-month-old infants could distinguish syllable sequences using statistical relationships after only two minutes of exposure to continuous artificial speech. This demonstrated a mechanism for finding candidate word boundaries, not acquisition of a complete vocabulary or grammar. Evidence of statistical learning with visual sequences also supports investigating domain-general learning alongside language-specific representations. Studies of multilingual processing similarly examine how linguistic knowledge interacts with selection mechanisms: Viorica Marian and Michael Spivey’s 2003 eye-tracking experiments found competition from objects whose names resembled a spoken target in either of a bilingual participant’s languages. Cross-language activation varied with language background and experimental conditions, rather than behaving as an invariably isolated second lexicon. (pubmed.ncbi.nlm.nih.gov)"
    },
    {
      "heading": "Mental Imagery and Spatial Cognition",
      "text": "Mental imagery involves sensory-like representations without the corresponding object or event being directly perceived: imagining a red door differs from seeing one because the door’s current sensory input is absent. Imagery is not exclusively visual; silently imagining a melody raises parallel questions about the representation of pitch, rhythm, and timbre. Andrea Halpern and Robert Zatorre’s 1999 study used positron-emission tomography to investigate imagery for familiar melodies, identifying contributions from auditory association and frontal regions. The central psychological questions concern which properties imagery preserves, which operations can transform it, and whether its resemblance to perception reflects shared mechanisms rather than subjective similarity alone. (bucknell.elsevierpure.com)\n\nIn Roger Shepard and Jacqueline Metzler’s 1971 mental-rotation experiment, participants judged whether two perspective drawings depicted the same three-dimensional block structure or different, mirror-related structures. Each participant completed 1,600 comparisons; the same-object pairs included angular disparities from \\(0^\\circ\\) to \\(180^\\circ\\), in \\(20^\\circ\\) steps. Response time increased approximately linearly with angular disparity, both for rotations within the picture plane and for rotations in depth. This pattern supported an analogue transformation account: larger imagined rotations require more processing time. Nevertheless, a linear response-time function constrains the sequence or duration of processing; it does not uniquely establish that a literal internal picture rotates, nor that every participant uses an identical strategy. (facultypsy.hope.edu)\n\nImage-scanning experiments examine a corresponding relationship between represented distance and processing time. In 1978, Stephen Kosslyn, Thomas Ball, and Brian Reiser asked participants to inspect remembered images, including a learned map, and mentally scan between locations. Across four experiments, scanning farther took longer even when the amount of intervening material was controlled; subjectively larger images also required longer scanning times. Importantly, distance did not affect response time when participants were not instructed to base their judgments on image inspection. The results therefore demonstrate distance-sensitive processing under particular task conditions, rather than an obligatory scanning mechanism whenever spatial information is remembered. (pubmed.ncbi.nlm.nih.gov)\n\nThese findings helped define the debate between depictive and propositional explanations. Depictive accounts, associated with Kosslyn, propose representations whose spatial organization preserves relationships such as relative position and distance; “depictive” need not mean a miniature photograph viewed by an internal observer. Propositional accounts, associated with Zenon Pylyshyn, instead emphasize structured descriptions—for example, above(triangle, square)—whose representational format need not itself be spatial. Pylyshyn argued that knowledge about how seeing and movement normally unfold could produce scanning effects through task-dependent expectations. Conversely, Kosslyn and William Thompson’s 2003 analysis found systematic, task-related variation in the recruitment of early visual cortex during imagery. Behavioral timing and neural overlap consequently constrain explanations, but neither alone supplies a complete account of representational format, image construction, and task strategy. (ruccs.rutgers.edu)\n\nSpatial cognition also requires reference frames. An egocentric description locates something relative to the observer—“to my left”—whereas an allocentric description relates locations to other environmental features—“beside the library.” Turning changes the first relationship without necessarily changing the second. Navigation combines such representations with landmarks and self-motion information: path integration estimates changes in position and heading from movement, while environmental cues help correct accumulated error. Remembering a sequence of turns can support a familiar route without providing sufficiently flexible spatial knowledge to identify a shortcut. Egocentric and allocentric information can operate together rather than constituting mutually exclusive navigation systems. (pubmed.ncbi.nlm.nih.gov)\n\nA cognitive map is an internal representation of environmental relationships, not an external cartographic display or necessarily a geometrically exact mental copy. Reading a visible map permits repeated inspection; navigating from memory requires retrieving and updating spatial relationships. Barbara Tversky’s 1981 experiments showed systematic distortions in remembered maps and environments: people tended to align separate geographical figures and rotate their remembered axes toward salient horizontal, vertical, or cardinal directions. These errors occurred even after explicit warnings. Spatial memory thus reflects organizational processes that make locations easier to encode and infer, while sometimes altering the directions or alignments subsequently reported. (sciencedirect.com)"
    },
    {
      "heading": "Reasoning, Problem Solving, and Creativity",
      "text": "Deductive reasoning concerns conclusions that necessarily follow from premises; inductive reasoning extends beyond the premises to conclusions that remain revisable; and analogical reasoning transfers relationships from a familiar situation to a less familiar one. Logical validity is a property of an argument, not a description of the psychological process producing an answer. A valid deduction can have a factually false conclusion if its premises are false, whereas a believable conclusion need not follow logically. Jonathan Evans, Julie Barston, and Paul Pollard’s 1983 syllogism experiments demonstrated this distinction experimentally: participants’ acceptance of conclusions depended on both validity and believability. Correct answers alone therefore do not establish that a person applied formal rules. (pubmed.ncbi.nlm.nih.gov)\n\nProblem solving depends on representing the initial situation, the goal, permissible actions, and constraints. Allen Newell and Herbert A. Simon described solving as search through a problem space: actions, or operators, transform one represented state into another. In means–ends analysis, a solver identifies a difference between the current state and the goal, selects an operator that could reduce it, and establishes subgoals when that operator’s prerequisites are unmet. Heuristics restrict the search without guaranteeing success. For example, selecting whichever move appears to bring the goal closer can fail when a solution requires temporarily increasing that distance. The psychological question is not simply whether a procedure finds an answer, but whether its sequence of operations resembles human choices, errors, and intermediate steps. (iiif.library.cmu.edu)\n\nInsight problems examine changes in representation rather than only continued search within an unchanged representation. In Karl Duncker’s candle task, reported in On Problem-Solving (1945), participants must mount a candle using materials that include tacks and a small box. The solution uses the box as a support, rather than merely as a container. Functional fixedness denotes the difficulty of recognizing such an alternative use. (cir.nii.ac.jp) Mental set instead concerns persistence with a previously successful procedure. In Abraham Luchins’s 1942 water-jar experiments, repeated problems encouraged the formula \\(B-A-2C\\): jars holding 127, 21, and 3 units yield 100 units by subtracting the smaller quantities from the largest. Participants could persist with this procedure on later problems permitting shorter solutions. Thus, prior success can constrain the alternatives considered, even when the available objects and operations are understood. (tandfonline.com)\n\nExpertise changes both representation and search. In four experiments published in 1981, Michelene Chi, Paul Feltovich, and Robert Glaser found that physics experts categorized problems principally by underlying principles, whereas novices relied more on literal features, such as the objects described. Such findings connect successful search to organized domain knowledge rather than to an unrestricted, general problem-solving capacity. (onlinelibrary.wiley.com) Transfer nevertheless requires recognizing when stored knowledge applies. Mary Gick and Keith Holyoak’s 1980 experiments used a military story as an analogy for Duncker’s radiation problem: dividing forces among converging routes corresponds to directing several weak rays toward one target. Transfer decreased substantially without a hint to use the earlier story, demonstrating that possessing a relevant analogy and spontaneously retrieving it are separable achievements. Expertise in one domain consequently does not itself establish competence in an unrelated one. (sciencedirect.com)\n\nCreativity involves producing possibilities and evaluating their novelty, usefulness, and fit to constraints. Ronald Finke, Thomas Ward, and Steven Smith’s 1992 Geneplore framework distinguishes generative processes from exploratory processes, with repeated movement between them rather than a compulsory linear sequence. Generation can combine remembered components into provisional structures; exploration interprets these structures, discovers possible functions, and tests their limitations. In imagery-based invention tasks, participants mentally combine simple forms and then explore potential uses for the resulting objects. These experiments operationalize creative activity as transformations and evaluations of representations, rather than treating an unusual answer or a subjective “aha” experience as a complete explanation of its origin. (direct.mit.edu)"
    },
    {
      "heading": "Judgment and Decision Making",
      "text": "Judgments estimate uncertain quantities or events, such as the probability that a candidate will win an election; decisions select actions, such as which candidate to support. Probability theory supplies standards for coherent judgments, including updating beliefs with Bayes’ theorem. Expected utility theory supplies a comparison standard for choices: an action \\(a\\) is evaluated by \\(EU(a)=\\sum_i P(s_i\\mid a)u(x_i)\\), where \\(s_i\\) denotes a possible state, \\(x_i\\) its outcome, and \\(u\\) the decision maker’s utility. Utility need not equal monetary value, so preferring a smaller certain payment to a gamble with a larger expected monetary value does not itself demonstrate inconsistency. These normative standards specify relationships among beliefs and preferences, rather than necessarily describing the psychological operations producing them. (stanford.edu)\n\nHerbert A. Simon challenged models that assume decision makers can identify every alternative, calculate its consequences, and select a global optimum. His 1955 paper “A Behavioral Model of Rational Choice” developed an account of choice constrained by available information and computational capacity, central to bounded rationality. In satisficing, alternatives are searched sequentially and evaluated against an aspiration level; search can stop when an acceptable option appears, without establishing that it is the best available. Aspiration levels can change with experience and the difficulty of finding satisfactory alternatives. This approach makes search order, stopping rules, and environmental structure explanatory variables, rather than treating every departure from maximization as an error. (doi.org)\n\nAmos Tversky and Daniel Kahneman’s 1974 analysis identified heuristics that simplify uncertain judgments. In an anchoring task, participants first compared the percentage of United Nations members that were African countries with a number produced by a wheel, then estimated the percentage. Starting numbers of 10 and 65 yielded median estimates of 25% and 45%, respectively. Availability concerns assessing frequency or probability through ease of retrieval: participants overestimated the relative frequency of English words beginning with certain consonants, including r, compared with words containing them in third position, where those consonants were actually more frequent. Representativeness concerns judging probability through resemblance: stereotypical personality descriptions influenced occupational judgments, sometimes at the expense of occupational base rates. These tasks distinguish arbitrary starting information, retrievability, and similarity as possible influences on judgment. (cebma.org)\n\nChoices also depend on how outcomes are described. In Tversky and Kahneman’s 1981 disease-outbreak problem, 152 participants chose between certainly saving 200 of 600 people and a gamble offering a one-third chance of saving everyone and a two-thirds chance of saving nobody; 72% selected the certain outcome. In a separate group of 155, equivalent outcomes were described as deaths: 400 deaths for certain, or a one-third chance of no deaths and a two-thirds chance of 600 deaths. Only 22% selected the certain outcome. This framing effect demonstrated a between-group reversal despite unchanged numerical consequences; it did not establish that every individual would reverse preference. The authors related such effects to evaluating outcomes relative to reference points as gains or losses. (stanford.edu)\n\nDual-process theories offer a family of explanations rather than identifying two literal, anatomically discrete brain systems. Jonathan Evans and Keith Stanovich’s 2013 formulation distinguished autonomous Type 1 processing from Type 2 processing requiring working-memory resources and hypothetical thinking. Some accounts propose that deliberation intervenes on an initial intuitive response; others allow concurrent, competing processes. Speed, emotionality, and correctness do not define the distinction: an intuitive answer can be correct, and effortful reasoning can support an incorrect conclusion. (krigolsonlab.com)\n\nTask descriptions can also change the computations needed for an answer. In Gerd Gigerenzer and Ulrich Hoffrage’s 1995 experiments, presenting Bayesian problems as naturally sampled frequencies—for example, nested counts rather than separate percentages—increased responses classified as using Bayesian algorithms. Across 15 problems, standard frequency presentations elicited 46% Bayesian solutions, compared with 16% for standard probability presentations. Their analysis connected this improvement to simpler computations. Such findings make information format part of the explanation of apparent bias, rather than supporting an unrestricted claim that people either possess or lack probabilistic competence. (pages.ucsd.edu)"
    },
    {
      "heading": "Metacognition and Conscious Access",
      "text": "Metacognition comprises monitoring cognition and controlling it in light of that monitoring. In Thomas O. Nelson and Louis Narens’s 1990 framework, information about an ongoing “object-level” activity informs a “meta-level” representation; control operates in the reverse direction, changing the activity. Monitoring includes retrospective confidence judgments about an answer and prospective judgments of learning (JOLs), which estimate the likelihood of remembering material later. Control includes choosing what to restudy, allocating study time, changing retrieval strategies, and deciding when to stop searching memory. The distinction is functional: estimating that a word pair is poorly learned is monitoring; selecting it for additional study is control. Accurate monitoring does not by itself establish effective control, because the resulting decisions must also be measured. (imbs.uci.edu)\n\nSubjective confidence, objective accuracy, and calibration describe different properties of performance. Accuracy concerns whether answers are correct; confidence concerns how likely participants believe those answers are to be correct. Calibration concerns their correspondence: among answers assigned 80% confidence, approximately 80% should be correct for that confidence category to be well calibrated. Metacognitive sensitivity instead concerns whether confidence distinguishes correct from incorrect responses across trials. Someone can systematically overestimate performance while still assigning higher confidence to correct answers than to errors. Stephen Fleming and Hakwan Lau’s 2014 methodological review therefore distinguished confidence bias, sensitivity, and efficiency, the latter assessing metacognitive sensitivity relative to first-order task performance. A high average confidence rating is not, by itself, evidence of accurate self-monitoring. (frontiersin.org)\n\nMonitoring relies on cues that need not predict durable knowledge. In Matthew Rhodes and Alan Castel’s 2008 experiments, participants gave higher JOLs to words printed in larger fonts, although font size showed little relationship with subsequent recall. Making large-print words harder to read eliminated the judgment difference, supporting a contribution from processing fluency rather than memory strength alone. Conversely, monitoring can become more diagnostic when the judgment situation resembles the eventual test. Nelson and John Dunlosky’s 1991 experiments found that delayed, cue-only JOLs predicted subsequent paired-associate recall substantially better than immediate judgments. Such results distinguish the accessibility of information during study from its later retrievability; neither ease of processing nor immediate familiarity guarantees accurate knowledge. (castel.psych.ucla.edu)\n\nResearch on conscious access separates awareness of a stimulus, assessment of one’s own performance, and the production of a report. Reporting how clearly a shape appeared asks about perceptual experience; reporting confidence that its orientation was correctly identified asks about decision accuracy. These questions are not interchangeable. The four-level Perceptual Awareness Scale, introduced by Thomas Ramsøy and Morten Overgaard in 2004, distinguishes no experience, a brief glimpse, almost clear experience, and clear experience rather than imposing a binary seen–unseen judgment. Different reporting methods can reveal different patterns of access, making the wording and scale of an awareness judgment part of the experimental measurement rather than a neutral window onto experience. (pmc.ncbi.nlm.nih.gov)\n\nProcessing without reportable awareness is investigated by combining direct visibility measures with indirect behavioral effects. In Lionel Naccache and Stanislas Dehaene’s 2001 numerical-priming experiments, participants classified visible numbers as smaller or larger than five. Masked preceding numbers facilitated responses when they implied the same classification, including primes never presented as visible targets. This supported semantic processing beyond learned associations between particular visible stimuli and responses. Nevertheless, failure to report a stimulus is not sufficient by itself to establish absence of experience: report thresholds and measurement sensitivity matter. Introspective explanations of why a response occurred also differ from reports of experienced content, a distinction emphasized by Richard Nisbett and Timothy Wilson in 1977. “No-report” paradigms address a further problem by reducing explicit reporting demands, but must still validate the indirect indicators used to infer experience. (sciencedirect.com)"
    },
    {
      "heading": "Brain–Cognition Relationships",
      "text": "Cognitive neuroscience connects functional accounts of perception, memory, language, and other cognitive processes with the organization and activity of neural systems. A functional distinction—such as maintaining information versus retrieving it—does not imply that each operation occupies a separate, self-contained brain region. An area may contribute to several tasks, while one task depends on interactions among multiple areas and their connections. Brain–behavior explanations therefore consider both regional specialization and network organization: damage can impair performance by destroying local tissue, interrupting communication, or altering activity elsewhere in the network. Anatomical localization constrains cognitive explanations without replacing the need to specify what information is processed and how. (pmc.ncbi.nlm.nih.gov)\n\nEvidence from brain injury is especially informative when abilities are differentially affected. In 1957, William Scoville and Brenda Milner described severe difficulty forming lasting new memories after bilateral medial temporal surgery in patients including H.M., despite relatively preserved general intelligence and perceptual abilities. Such a single dissociation suggests that the affected and preserved performances do not depend identically on the damaged system, but unequal task difficulty can provide an alternative explanation. A double dissociation strengthens the inference: one patient performs poorly on task A but relatively well on B, while another shows the reverse pattern. Nevertheless, this does not by itself prove that two independent anatomical modules exist; conclusions depend on task sensitivity, assumptions about processing, and possible compensatory strategies. Naturally occurring lesions also differ in extent and location, and post-injury reorganization complicates comparisons with intact brains. (pmc.ncbi.nlm.nih.gov)\n\nElectroencephalography (EEG) records voltage differences between electrodes, usually placed on the scalp, reflecting the summed electrical activity of neuronal populations. Its millisecond-scale temporal resolution helps establish when experimental conditions begin to produce different responses. Event-related potentials (ERPs) are obtained by aligning EEG segments to an event, such as word onset, and commonly averaging across trials to reveal time-locked activity. In 1980, Marta Kutas and Steven Hillyard reported that semantically inappropriate sentence endings elicited a negative-going response peaking approximately 400 milliseconds after word presentation—the N400—whereas unusually large printed words produced a different, positive response. This distinguished responses to semantic and physical deviations in time. Scalp distributions, however, do not directly identify their neural generators: electrical fields spread through head tissues, and different source configurations can produce similar recordings. (nobaproject.com)\n\nFunctional magnetic resonance imaging (fMRI) provides complementary spatial information, commonly at a scale of millimeters. Its blood-oxygen-level-dependent (BOLD) signal reflects changes in blood oxygenation associated with neural activity, rather than directly recording neuronal firing. The vascular response unfolds over seconds, limiting the separation of rapidly successive cognitive events even when images are acquired frequently. In 2001, Nikos Logothetis and colleagues simultaneously recorded electrophysiological activity and fMRI responses in monkey visual cortex; local field potentials predicted BOLD responses better than multi-unit spiking under their experimental conditions. This finding linked the imaging signal particularly to local input and processing rather than simply neuronal output. A task-related BOLD increase consequently identifies an association under a specified experimental contrast, not proof that the region exclusively performs—or is necessary for—the proposed mental operation. (nobaproject.com)\n\nTranscranial magnetic stimulation (TMS) introduces a perturbation rather than merely recording activity: a rapidly changing magnetic field induces an electric field that alters neural activity. Varying stimulation location and timing while measuring performance can test a region’s contribution during a particular processing interval. Electrical stimulation methods likewise manipulate activity, but their effects depend on current distribution and neural state. A behavioral change supports causal inference about the intervention; attributing it specifically to a targeted cognitive mechanism additionally requires controls for stimulation sounds, scalp sensations, unintended stimulation, and effects propagated through connected regions. Conversely, a null result does not establish that the target is unnecessary, because stimulation may have been insufficient or poorly timed. (pubmed.ncbi.nlm.nih.gov)\n\nMultimethod convergence addresses these complementary limitations. Lesion evidence can constrain which structures are required for normal performance; EEG and ERPs constrain processing time; fMRI identifies spatially distributed task-related responses; and stimulation tests the consequences of perturbing selected parts of the system. Agreement across methods is strongest when they test the same explicitly defined cognitive operation. Evidence that disrupting a structure impairs a task supports its necessity under those conditions, not its sufficiency to generate the ability or its exclusive ownership of that function. (pmc.ncbi.nlm.nih.gov)"
    },
    {
      "heading": "Development, Aging, and Sociocultural Context",
      "text": "Jean Piaget’s account of cognitive development proposes qualitative reorganizations of thought: sensorimotor activity gives way to symbolic representation, concrete operations, and eventually formal reasoning. In conservation tasks, a child may judge that pouring liquid into a taller, narrower glass increases its quantity; understanding conservation requires coordinating appearance with reversible transformations. Information-processing approaches instead analyze changes in processing efficiency, memory, knowledge, and strategy selection, allowing different components to develop at different rates. Lev Vygotsky’s sociocultural approach emphasizes the mediation of thinking by language, cultural tools, and interaction. Its “zone of proximal development” distinguishes independent performance from what becomes possible with assistance. These approaches therefore locate developmental change at different explanatory levels rather than simply offering competing age schedules. (nobaproject.com)\n\nMaturation and experience interact rather than contribute independently to a fixed timetable. Increasing attentional control can support more complex strategies, while instruction and practice change which strategies are available and how efficiently they operate. Apparent discontinuities in successful performance can consequently emerge from gradual changes in several underlying processes; conversely, children of the same age may perform differently across familiar and unfamiliar tasks. (nobaproject.com) Robert Siegler and Eric Jenkins’s 1989 research distinguished discovering an arithmetic strategy from generalizing it. Trial-by-trial observation made it possible to identify a strategy’s first use without assuming that discovery immediately produced consistent application across problems. (routledge.com)\n\nAdult cognitive change is likewise asynchronous. Joshua Hartshorne and Laura Germine’s 2015 investigation combined results from 48,537 online participants with standardized-test normative data. Different abilities showed different age profiles: some peaked around late adolescence, others in early adulthood, and still others in the forties or later. Thus, processing speed, memory performance, and accumulated verbal knowledge cannot be represented adequately by one age of “peak cognition.” These are population-level performance curves, not schedules that determine an individual’s trajectory; the study also found that the timing of vocabulary peaks differed across historical datasets. (pubmed.ncbi.nlm.nih.gov)\n\nThe interpretation of aging depends on research design. Cross-sectional comparisons measure different age groups at one time, so age differences can also reflect cohort differences in schooling and experience. Longitudinal studies follow the same individuals, but repeated exposure to tests can improve scores and obscure underlying decline. Timothy Salthouse’s 2009 study used three approaches to estimating retest effects and concluded that prior testing substantially contributed to discrepancies between cross-sectional and longitudinal patterns. The distinction is between differences among people of different ages and changes within people over time: neither design, without attention to its confounds, supplies a direct measure of a universal aging process. (pubmed.ncbi.nlm.nih.gov)\n\nLiteracy and schooling are experiences that can alter the processes recruited by cognitive tasks, rather than merely reveal pre-existing ability. In 2010, Stanislas Dehaene and colleagues compared 10 adults who could not read, 22 who acquired literacy in adulthood, and 31 who learned in childhood. Literacy was associated with stronger responses to writing in left fusiform cortex and changes in visual and spoken-language processing; many differences were also present among adult learners. More directly addressing change over time, a 2017 longitudinal study found altered visual-system connectivity after six months of adult literacy training, with connectivity changes related to gains in decoding skill. These findings demonstrate experience-related plasticity without making literacy a measure of general cognitive worth. (pubmed.ncbi.nlm.nih.gov)\n\nCultural practices also shape the communicative setting in which competence is expressed. Barbara Rogoff and colleagues’ 1993 study observed 14 toddlers aged 12–24 months and their caregivers in each of four communities in Guatemala, India, Turkey, and the United States. Their analysis examined both shared processes of guided participation and differences in how adults structured activities and children observed or participated. Such evidence makes the organization of learning itself a research variable. Performance elicited through explicit questioning or adult-directed demonstration must therefore be interpreted in relation to participants’ experience of those interactional practices, rather than treated automatically as a culture-independent index of developmental level. (pubmed.ncbi.nlm.nih.gov)"
    },
    {
      "heading": "Applications in Education, Design, and Law",
      "text": "In education, cognitive research makes instructional sequencing an experimentally testable question: does presenting a complete solution before independent practice change learning time, errors, or transfer? John Sweller and Graham A. Cooper’s 1985 algebra experiments found that studying worked examples reduced instructional time and improved subsequent solution speed and accuracy for problems sharing the examples’ structure. Their proposed mechanism was that examples reduce the search for solution steps, allowing learners to acquire schemas for recognizing and solving particular problem types. The structural restriction matters: improved performance on similar equations did not establish a general improvement in mathematical reasoning. This work connects instructional demands to the knowledge learners already possess, rather than treating task difficulty as an invariant property of the material. (tandfonline.com)\n\nAssessment can also change the knowledge it measures. Henry L. Roediger III and Jeffrey D. Karpicke’s 2006 experiments showed that recalling prose material during learning improved retention after a week relative to additional reading. Thus, an assessment schedule can function as part of instruction, not merely as a final measurement. Feedback introduces further design variables. In two experiments published in 2007, Andrew C. Butler, Karpicke, and Roediger independently manipulated when feedback appeared and whether participants saw the answer or continued choosing until correct. Delayed feedback produced better subsequent cued recall than immediate feedback, whereas the two feedback formats produced equivalent performance. These findings distinguish feedback timing from response format and make delayed retention, rather than immediate success alone, an educational outcome. (psychologicalscience.org)\n\nHuman factors applies cognitive findings to interfaces, displays, warnings, and operational environments. Evaluation concerns what users notice, understand, and do during specified tasks, not simply whether a display presents technically accurate information. The US Federal Aviation Administration’s 2013 guide to multifunction displays explicitly separated human-factors evaluation from equipment functionality and operational accuracy. Warning systems illustrate the importance of this distinction: a 2007 FAA analysis of air-traffic-control safety alerts identified frequent nuisance alerts as a source of distraction, additional workload, and reduced trust in automation. Increasing an alert’s loudness could therefore fail to improve responses if controllers continued to encounter many unhelpful alerts. The relevant design problem includes alert frequency, priority, suppression controls, and interference with ongoing work. (faa.gov)\n\nIn law, research on eyewitness identification and testimony examines how questioning and identification procedures affect evidence. Elizabeth F. Loftus and John C. Palmer’s 1974 experiments showed that wording questions about filmed collisions changed reported speed. In a second experiment involving 150 participants, those questioned using “smashed” were more likely than those questioned using “hit” to report broken glass a week later, although none appeared in the film. The results connected post-event language with later recollection; they did not establish that every discrepancy in testimony reflects a changed memory rather than a response bias. (gwern.net)\n\nProcedural applications distinguish conditions during the witnessed event from influences introduced during investigation. The National Research Council’s 2014 report Identifying the Culprit recommended double-blind lineup administration, standardized witness instructions, recording confidence at the initial identification, and video recording the procedure. In double-blind administration, the administrator does not know which lineup member is the suspect, limiting opportunities for inadvertent cues. Recording confidence immediately preserves a statement made before subsequent feedback can inflate it. These were research-informed recommendations for handling evidence, not a claim that any procedure guarantees an accurate identification. (nationalacademies.org)\n\nCognitive modeling provides another application in human–computer interaction. Stuart K. Card, Thomas P. Moran, and Allen Newell’s 1980 keystroke-level model predicted the time an expert user would require for a specified routine task by summing estimated durations for keystrokes, pointing, mental preparation, and system responses. It thereby made alternative interaction sequences quantitatively comparable. Its scope was deliberately narrower than complete user behavior: predicting skilled execution time does not explain how novices discover a command or recover from misunderstanding. This illustrates the distinction from artificial-intelligence systems evaluated primarily by successful task completion: a psychological model must additionally account for human processing constraints and observed performance patterns, rather than merely produce a correct result. (researchgate.net)"
    },
    {
      "heading": "Reproducibility, Generalizability, and Explanatory Limits",
      "text": "Direct replication repeats a study’s procedures as closely as practicable with new observations; conceptual replication changes the task or operationalization while testing the same theoretical claim. Direct replication primarily assesses the dependability of a finding under specified conditions, whereas conceptual replication also tests whether its interpretation extends beyond one procedure. In 2015, the Open Science Collaboration reported replications of 100 studies published in three psychology journals in 2008. Statistically significant results occurred in 97% of the original studies but 36% of the replications, and replication effect sizes averaged approximately half the original magnitudes. These figures describe a selected, mixed sample of psychological research, not the proportion of all cognitive findings that are true. Reproducibility therefore concerns effect magnitudes and uncertainty as well as whether a significance threshold is crossed. (nobaproject.com)\n\nStatistical power is the probability of rejecting a false null hypothesis for a specified effect and design. Low-powered studies frequently miss genuine effects; when publication favors significant outcomes, the published subset can overestimate their magnitude. Publication bias also leaves null findings less visible, while undisclosed choices among analyses can increase false-positive results. Transparent reporting makes these decisions inspectable: shared stimuli, task software, anonymized data, and analysis code allow procedural checking and computational reproduction. Preregistration records hypotheses and analysis plans before outcomes are examined, distinguishing planned tests from exploratory analyses. Registered Reports add results-blind peer review and an in-principle publication decision before the research outcomes are known, reducing incentives to publish only positive findings. (nobaproject.com)\n\nA replicable experimental contrast need not reliably measure differences between people. Craig Hedge, Georgina Powell, and Petroc Sumner demonstrated this “reliability paradox” in research published in 2018, examining seven established cognitive tasks across three studies. A task can consistently produce interference at the group level while poorly preserving individuals’ rank ordering across testing sessions. When participants have similar true interference effects, measurement noise can dominate the small between-person differences. Consequently, a robust mean effect does not automatically justify correlating individual interference scores with other abilities or neural measures. Nor does replication uniquely establish a theory: reproducing an effect confirms an observation that several mechanisms may predict. (pmc.ncbi.nlm.nih.gov)\n\nGeneralization depends on the populations, materials, and situations sampled. Joseph Henrich, Steven Heine, and Ara Norenzayan’s 2010 review documented cross-population variation in domains including visual perception, spatial reasoning, and categorization, challenging universal claims based predominantly on Western, educated, industrialized, rich, democratic populations. Participant diversity is only one dimension: effects obtained with a particular word list, image set, or laboratory task may not extend to other materials. Ecological validity concerns relevance to activities outside the testing arrangement, rather than merely whether an experiment looks realistic. Simplified tasks isolate variables, but extrapolation to everyday activity requires evidence about the changed demands and contexts; adding participants alone does not establish generality across untested tasks or stimuli. (doi.org)\n\nExplanatory limits remain even when measurements replicate and generalize. Competing computational models can fit the same observations, making goodness of fit insufficient to identify the underlying process. Model-recovery tests address this problem by generating artificial data from known candidate models and checking whether the comparison procedure identifies their source; parameter recovery similarly tests whether fitted parameters recover known generating values. Integration across perception, memory, language, and action imposes further constraints beyond success on isolated tasks. Olivia Guest and Andrea Martin argued in 2021 that computational implementation strengthens theory building by forcing assumptions to become explicit. Nevertheless, computational success is not itself psychological plausibility: a model must be evaluated against human behavioral patterns and its intended explanatory commitments, not simply whether it produces correct task answers. (elifesciences.org)"
    }
  ],
  "links": [
    "psychology",
    "cognitive-science",
    "neuroscience",
    "linguistics",
    "artificial-intelligence",
    "language",
    "philosophy",
    "anthropology",
    "brain",
    "clinical-psychology",
    "cognitive-behavioral-therapy",
    "information-theory",
    "herbert-a-simon",
    "experimental-design",
    "randomization",
    "confounding",
    "correlation",
    "causal-inference",
    "connectionism",
    "artificial-neural-network",
    "bayesian-inference",
    "bayes-theorem",
    "reading-comprehension",
    "knowledge",
    "inductive-reasoning",
    "phonology",
    "syntax",
    "semantics",
    "orthography",
    "reading",
    "language-acquisition",
    "multilingualism",
    "deductive-reasoning",
    "logical-validity",
    "heuristic",
    "probability",
    "expected-value",
    "bounded-rationality",
    "cognitive-bias",
    "consciousness",
    "neuroplasticity",
    "electroencephalography",
    "magnetic-resonance-imaging",
    "literacy",
    "education",
    "automation",
    "law",
    "reproducibility",
    "statistical-power",
    "peer-review"
  ],
  "sources": [
    {
      "title": "Cognitive Psychology Explores Our Mental Processes",
      "url": "https://www.apa.org/action/science/brain-science"
    },
    {
      "title": "APA Dictionary of Psychology: Cognitive Psychology",
      "url": "https://dictionary.apa.org/cognitive-psychology?cc_opt_out=true&utm_campaign%5B0%5D%5B0%5D=132064270743&utm_campaign%5B0%5D%5B1%5D=139166491354&utm_campaign%5B1%5D%5B0%5D=132064270743&utm_campaign%5B1%5D%5B1%5D=139166491354&utm_content%5B0%5D%5B0%5D=582218976168&utm_content%5B0%5D%5B1%5D=595535453738&utm_content%5B1%5D%5B0%5D=582218976168&utm_content%5B1%5D%5B1%5D=595535453738&utm_device%5B0%5D%5B0%5D=cutm_source%3Dgoogleutm_medium%3Dcpcutm_campaign%3D136246792271utm_content%3D582218976177utm_term%3Dgrow+therapygclid%3DCjwKCAiAk--dBhABEiwAchIwkeoLZ_jSxF9u3aXP2z2dns8BpX21ZiAJmbw9HgeOTz1UqI4gCs6UdBoCpdsQAvD_BwE&utm_device%5B0%5D%5B1%5D=c&utm_device%5B1%5D%5B0%5D=cutm_source%3Dgoogleutm_medium%3Dcpcutm_campaign%3D136246792271utm_content%3D582218976177utm_term%3Dgrow+therapygclid%3DCjwKCAiAk--dBhABEiwAchIwkeoLZ_jSxF9u3aXP2z2dns8BpX21ZiAJmbw9HgeOTz1UqI4gCs6UdBoCpdsQAvD_BwE&utm_device%5B1%5D%5B1%5D=c&utm_device%5B2%5D%5B0%5D=cutm_source%3Dgoogleutm_medium%3Dcpcutm_campaign%3D136246792271utm_content%3D582218976177utm_term%3Dgrow+therapygclid%3DCjwKCAiAk--dBhABEiwAchIwkeoLZ_jSxF9u3aXP2z2dns8BpX21ZiAJmbw9HgeOTz1UqI4gCs6UdBoCpdsQAvD_BwE&utm_device%5B2%5D%5B1%5D=c&utm_device%5B3%5D=cutm_source%3Dgoogleutm_medium%3Dcpcutm_campaign%3D136246792271utm_content%3D582218976177utm_term%3Dgrow+therapygclid%3DCjwKCAiAk--dBhABEiwAchIwkeoLZ_jSxF9u3aXP2z2dns8BpX21ZiAJmbw9HgeOTz1UqI4gCs6UdBoCpdsQAvD_BwE&utm_device%5B4%5D=c&utm_device%5B5%5D=c&utm_medium%5B0%5D%5B0%5D=cpc&utm_medium%5B0%5D%5B1%5D=cpc&utm_medium%5B1%5D%5B0%5D=cpc&utm_medium%5B1%5D%5B1%5D=cpc&utm_source%5B0%5D%5B0%5D=google&utm_source%5B0%5D%5B1%5D=google&utm_source%5B1%5D%5B0%5D=google&utm_source%5B1%5D%5B1%5D=google&utm_term%5B0%5D%5B0%5D=grow+therapy&utm_term%5B0%5D%5B1%5D=melissa+galica&utm_term%5B1%5D%5B0%5D=grow+therapy&utm_term%5B1%5D%5B1%5D=melissa+galica"
    },
    {
      "title": "High-speed scanning in human memory",
      "url": "https://pubmed.ncbi.nlm.nih.gov/5939936/"
    },
    {
      "title": "Home — Cognitive Science Society",
      "url": "https://cognitivesciencesociety.org/"
    },
    {
      "title": "Cognitive computational neuroscience",
      "url": "https://arxiv.org/abs/1807.11819"
    },
    {
      "title": "Clinical Psychology Solves Complex Human Problems",
      "url": "https://www.apa.org/education-career/guide/subfields/clinical"
    },
    {
      "title": "What is Cognitive Behavioral Therapy?",
      "url": "https://www.apa.org/ptsd-guideline/patients-and-families/cognitive-behavioral?os=av%2F"
    },
    {
      "title": "Geschichte des Instituts",
      "url": "https://www.lw.uni-leipzig.de/wilhelm-wundt-institut-fuer-psychologie/institutsgeschichte/geschichte-des-instituts"
    },
    {
      "title": "Gustav Theodor Fechners chronologisch-tabellarischer Lebenslauf",
      "url": "https://home.uni-leipzig.de/wundtbriefe/wwcd/fechlife/leben.htm"
    },
    {
      "title": "Memory: A Contribution to Experimental Psychology, Chapter VIII",
      "url": "https://psychclassics.yorku.ca/Ebbinghaus/memory8.htm"
    },
    {
      "title": "Psychology as the Behaviorist Views It",
      "url": "https://psychclassics.yorku.ca/Watson/views.htm"
    },
    {
      "title": "History of Psychology",
      "url": "https://nobaproject.com/modules/history-of-psychology"
    },
    {
      "title": "Cognitive Maps in Rats and Men",
      "url": "https://psychclassics.yorku.ca/Tolman/Maps/maps"
    },
    {
      "title": "Laws of Organization in Perceptual Forms",
      "url": "https://psychclassics.yorku.ca/Wertheimer/Forms/forms.htm?source=post_page---------------------------"
    },
    {
      "title": "The Cognitive Revolution: A Historical Perspective",
      "url": "https://www.cs.princeton.edu/~rit/geo/Miller.pdf"
    },
    {
      "title": "Jerome Bruner",
      "url": "https://psychology.fas.harvard.edu/people/jerome-bruner"
    },
    {
      "title": "The Cognitive Revolution",
      "url": "https://news.harvard.edu/gazette/story/2011/10/the-cognitive-revolution/"
    },
    {
      "title": "The Cognitive Approach",
      "url": "https://stanford.edu/class/psych205/papers/Neisser-1967_ch1.pdf"
    },
    {
      "title": "Psychology Turns Cognitive",
      "url": "https://mechanism.ucsd.edu/~bill/teaching/philpsych/cognitivepsych.pdf"
    },
    {
      "title": "Research Designs",
      "url": "https://nobaproject.com/modules/research-designs"
    },
    {
      "title": "Studies of Interference in Serial Verbal Reactions",
      "url": "https://psychclassics.yorku.ca/Stroop/"
    },
    {
      "title": "Within-subjects vs. Between-subjects Designs: Which to Use?",
      "url": "https://www.yorku.ca/mack/RN-Counterbalancing.html"
    },
    {
      "title": "Memory (Encoding, Storage, Retrieval)",
      "url": "https://nobaproject.com/modules/memory-encoding-storage-retrieval"
    },
    {
      "title": "Toward a model of eye movement control in reading.",
      "url": "https://pubmed.ncbi.nlm.nih.gov/9450374/"
    },
    {
      "title": "A maximum-likelihood procedure for estimating psychometric functions: Thresholds, slopes, and lapses of attention",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC3427362/"
    },
    {
      "title": "Slow and steady? Strategic adjustments in response caution are moderately reliable and correlate across tasks",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC6920044/"
    },
    {
      "title": "The reliability paradox: Why robust cognitive tasks do not produce reliable individual differences",
      "url": "https://pubmed.ncbi.nlm.nih.gov/28726177/"
    },
    {
      "title": "The reliability paradox: Why robust cognitive tasks do not produce reliable individual differences",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC5990556/"
    },
    {
      "title": "An Integrated Theory of the Mind",
      "url": "https://www.cs.utexas.edu/~dana/ACT-R.pdf"
    },
    {
      "title": "An Interactive Activation Model of Context Effects in Letter Perception: Part 1. An Account of Basic Findings",
      "url": "https://stanford.edu/~jlmcc/papers/McClellandRumelhart81.pdf"
    },
    {
      "title": "Learning representations by back-propagating errors",
      "url": "https://doi.org/10.1038/323533a0"
    },
    {
      "title": "Probabilistic models of cognition: exploring representations and inductive biases",
      "url": "https://digital.library.adelaide.edu.au/items/32a7b66a-79e9-415c-9843-da9ed12515eb"
    },
    {
      "title": "Sensation and Perception",
      "url": "https://nobaproject.com/textbooks/jack-arnal-introduction-to-psychology-the-full-noba-collection/modules/sensation-and-perception"
    },
    {
      "title": "Neural correlates of auditory stream segregation in primary auditory cortex of the awake monkey",
      "url": "https://pubmed.ncbi.nlm.nih.gov/11124464/"
    },
    {
      "title": "Recognition-by-Components: A Theory of Human Image Understanding",
      "url": "https://www.cs.utexas.edu/~grauman/courses/spring2007/395T/papers/Recognition-by-Components.PDF"
    },
    {
      "title": "Feed-Forward Segmentation of Figure-Ground and Assignment of Border-Ownership",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC2873306/"
    },
    {
      "title": "Depth perception from dynamic occlusion in motion parallax: Roles of expansion-compression versus accretion-deletion",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4521857/"
    },
    {
      "title": "Vision",
      "url": "https://nobaproject.com/modules/vision"
    },
    {
      "title": "The effects of contextual scenes on the identification of objects",
      "url": "https://link.springer.com/content/pdf/10.3758%2FBF03197524.pdf"
    },
    {
      "title": "Parts and wholes in face recognition",
      "url": "https://pubmed.ncbi.nlm.nih.gov/8316637/"
    },
    {
      "title": "Hearing lips and seeing voices",
      "url": "https://www.nature.com/articles/264746a0"
    },
    {
      "title": "Humans integrate visual and haptic information in a statistically optimal fashion",
      "url": "https://doi.org/10.1038/415429a"
    },
    {
      "title": "Attention",
      "url": "https://nobaproject.com/modules/attention"
    },
    {
      "title": "Involuntary covert orienting is contingent on attentional control settings",
      "url": "https://pubmed.ncbi.nlm.nih.gov/1431742/"
    },
    {
      "title": "A feature-integration theory of attention",
      "url": "https://cseweb.ucsd.edu/classes/fa09/cse258a/papers/treisman_gelade.pdf"
    },
    {
      "title": "The Temporal Dynamics of Visual Search: Evidence for Parallel Processing in Feature and Conjunction Searches",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC3313830/"
    },
    {
      "title": "Gorillas in our midst: sustained inattentional blindness for dynamic events",
      "url": "https://www.chabris.com/Simons1999.pdf"
    },
    {
      "title": "What are the differences between long-term, short-term, and working memory?",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC2657600/"
    },
    {
      "title": "The episodic buffer: a new component of working memory?",
      "url": "https://doi.org/10.1016/s1364-6613%2800%2901538-2"
    },
    {
      "title": "The magical number 4 in short-term memory: a reconsideration of mental storage capacity",
      "url": "https://pubmed.ncbi.nlm.nih.gov/11515286/?dopt=Abstract"
    },
    {
      "title": "Phonological recoding under articulatory suppression",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC5809545/"
    },
    {
      "title": "Individual Differences in Working Memory and Reading.",
      "url": "https://eric.ed.gov/?id=EJ230538"
    },
    {
      "title": "Why Does Working Memory Capacity Predict Variation in Reading Comprehension? On the Influence of Mind Wandering and Executive Attention",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC3387280/"
    },
    {
      "title": "Structure and function of declarative and nondeclarative memory systems",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/33639/"
    },
    {
      "title": "Forgetting and Amnesia",
      "url": "https://nobaproject.com/textbooks/constance-toffle-ph-d-new-textbook/modules/forgetting-and-amnesia"
    },
    {
      "title": "The Psychology and Neuroscience of Forgetting",
      "url": "https://cenl.ucsd.edu/Jclub/Wixted_2004.pdf"
    },
    {
      "title": "Source monitoring",
      "url": "https://memlab.yale.edu/sites/default/files/files/1993_Johnson_Hashtroudi_Lindsay_PsychBull.pdf"
    },
    {
      "title": "The Legacy of Patient H.M. for Neuroscience",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC2649674/"
    },
    {
      "title": "Learning Versus Performance",
      "url": "https://www.psychologicalscience.org/journals/perspectives/1745691615569000/"
    },
    {
      "title": "Factors Influencing Learning",
      "url": "https://nobaproject.com/modules/factors-influencing-learning"
    },
    {
      "title": "Where practice makes perfect in texture discrimination: evidence for primary visual cortex plasticity",
      "url": "https://pubmed.ncbi.nlm.nih.gov/2052578/"
    },
    {
      "title": "Human Neural Learning Depends on Reward Prediction Errors in the Blocking Paradigm",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC2637603/"
    },
    {
      "title": "Test-enhanced learning: taking memory tests improves long-term retention",
      "url": "https://pubmed.ncbi.nlm.nih.gov/16507066/"
    },
    {
      "title": "Feedback enhances the positive effects and reduces the negative effects of multiple-choice testing.",
      "url": "https://pubmed.ncbi.nlm.nih.gov/18491500/"
    },
    {
      "title": "Mathematical problem solving by analogy",
      "url": "https://pubmed.ncbi.nlm.nih.gov/1829473/"
    },
    {
      "title": "Categories and Concepts",
      "url": "https://nobaproject.com/modules/categories-and-concepts"
    },
    {
      "title": "Attention, similarity, and the identification-categorization relationship",
      "url": "https://pubmed.ncbi.nlm.nih.gov/2937873"
    },
    {
      "title": "Attention, Similarity, and the Identification-Categorization Relationship",
      "url": "https://psych.indiana.edu/documents/nosofsky-1986.pdf"
    },
    {
      "title": "Family Resemblances: Studies in the Internal Structure of Categories",
      "url": "https://worthylab.org/wp-content/uploads/2021/08/roschmervis1975_familyresemblance_cognitivepsychology.pdf"
    },
    {
      "title": "The role of theories in conceptual coherence",
      "url": "https://pubmed.ncbi.nlm.nih.gov/4023146/"
    },
    {
      "title": "The Role of Theories in Conceptual Coherence",
      "url": "https://worthylab.org/wp-content/uploads/2021/08/murphymedin1985_conceptualcoherence_psychologicalreview.pdf"
    },
    {
      "title": "The locus of knowledge effects in concept learning",
      "url": "https://pubmed.ncbi.nlm.nih.gov/8064251/"
    },
    {
      "title": "The TRACE Model of Speech Perception",
      "url": "https://web.stanford.edu/~jlmcc/papers/McClellandElman86.pdf"
    },
    {
      "title": "A theory of lexical access in speech production",
      "url": "https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/abs/theory-of-lexical-access-in-speech-production/7E4A98E8791AB85397761DAAB35288AA"
    },
    {
      "title": "Integration of Visual and Linguistic Information in Spoken Language Comprehension",
      "url": "https://shotam.github.io/LING611_papers/Tanenhaus_1995.pdf"
    },
    {
      "title": "Reading senseless sentences: brain potentials reflect semantic incongruity",
      "url": "https://pubmed.ncbi.nlm.nih.gov/7350657/"
    },
    {
      "title": "DRC: a dual route cascaded model of visual word recognition and reading aloud",
      "url": "https://pubmed.ncbi.nlm.nih.gov/11212628/"
    },
    {
      "title": "Statistical learning by 8-month-old infants",
      "url": "https://pubmed.ncbi.nlm.nih.gov/8943209/"
    },
    {
      "title": "Visual statistical learning in infancy: evidence for a domain general learning mechanism",
      "url": "https://pubmed.ncbi.nlm.nih.gov/11869728/"
    },
    {
      "title": "Competing activation in bilingual language processing: Within- and between-language competition",
      "url": "https://www.cambridge.org/core/journals/bilingualism-language-and-cognition/article/abs/competing-activation-in-bilingual-language-processing-within-and-betweenlanguage-competition/550B320E3F431CECFA8B2D60F37FB40B"
    },
    {
      "title": "When that tune runs through your head: A PET investigation of auditory imagery for familiar melodies",
      "url": "https://bucknell.elsevierpure.com/en/publications/when-that-tune-runs-through-your-head-a-pet-investigation-of-audi-2/"
    },
    {
      "title": "Mental Rotation of Three-Dimensional Objects",
      "url": "https://facultypsy.hope.edu/psychlabs/exp/rotate/readings/ShepardMetzler_1971.pdf"
    },
    {
      "title": "Visual images preserve metric spatial information: evidence from studies of image scanning",
      "url": "https://pubmed.ncbi.nlm.nih.gov/627850/"
    },
    {
      "title": "When is early visual cortex activated during visual mental imagery?",
      "url": "https://pubmed.ncbi.nlm.nih.gov/12956541/"
    },
    {
      "title": "Spatial memory: how egocentric and allocentric combine",
      "url": "https://pubmed.ncbi.nlm.nih.gov/17071127/"
    },
    {
      "title": "Distortions in memory for maps",
      "url": "https://www.tc.columbia.edu/faculty/bt2158/faculty-profile/files/1981_Tversky_Distortionsinmemoryformaps.pdf"
    },
    {
      "title": "On the conflict between logic and belief in syllogistic reasoning.",
      "url": "https://pubmed.ncbi.nlm.nih.gov/6621345/"
    },
    {
      "title": "HUMAN PROBLEM SOLVING",
      "url": "https://iiif.library.cmu.edu/file/Newell_box00018_fld01306_doc0001/Newell_box00018_fld01306_doc0001.pdf"
    },
    {
      "title": "On problem-solving.",
      "url": "https://cir.nii.ac.jp/crid/1364233269905221632"
    },
    {
      "title": "Problem Solving",
      "url": "https://www.yorku.ca/vgoel/reprints/problem_solving.pdf"
    },
    {
      "title": "When High WMC Promotes Mental Set: A Model of the Water Jar Task",
      "url": "https://escholarship.org/uc/item/8gz9d6m4"
    },
    {
      "title": "Satisficing: Integrating Two Traditions",
      "url": "https://pure.mpg.de/rest/items/item_3262700_13/component/file_3457640/content"
    },
    {
      "title": "Judgment under Uncertainty: Heuristics and Biases",
      "url": "https://cebma.org/assets/Uploads/Tversky_Kahneman_1974.pdf"
    },
    {
      "title": "The Framing of Decisions and the Psychology of Choice",
      "url": "https://stanford.edu/class/psych205/papers/Tversky-Kahneman-1981.pdf"
    },
    {
      "title": "Dual-Process Theories of Higher Cognition: Advancing the Debate",
      "url": "https://www.krigolsonlab.com/uploads/4/3/8/4/43848243/evans_and_stanovich_2013.pdf"
    },
    {
      "title": "How to Improve Bayesian Reasoning Without Instruction: Frequency Formats",
      "url": "https://pages.ucsd.edu/~scoulson/203/GG_How_1995.pdf"
    },
    {
      "title": "How to measure metacognition",
      "url": "https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2014.00443/full"
    },
    {
      "title": "Memory Predictions Are Influenced by Perceptual Information: Evidence for Metacognitive Illusions",
      "url": "https://castel.psych.ucla.edu/wp-content/uploads/sites/111/2016/11/RhodesCastelJEPG2008.pdf"
    },
    {
      "title": "Kinds of access: different methods for report reveal different kinds of metacognitive access",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC3318768/"
    },
    {
      "title": "No-report paradigms: Extracting the true neural correlates of consciousness",
      "url": "https://research.monash.edu/en/publications/no-report-paradigms-extracting-the-true-neural-correlates-of-cons/"
    },
    {
      "title": "Lesion mapping in neuropsychological research: A practical and conceptual guide",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11474248"
    },
    {
      "title": "Lesion studies in contemporary neuroscience",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC6712987/"
    },
    {
      "title": "LOSS OF RECENT MEMORY AFTER BILATERAL HIPPOCAMPAL LESIONS",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/497229/"
    },
    {
      "title": "Kable: Neuropsychological methods",
      "url": "https://web.stanford.edu/~knutson/nfc/kable12.pdf"
    },
    {
      "title": "Psychophysiological Methods in Neuroscience",
      "url": "https://nobaproject.com/modules/psychophysiological-methods-in-neuroscience"
    },
    {
      "title": "Reading Senseless Sentences: Brain Potentials Reflect Semantic Incongruity",
      "url": "https://pages.ucsd.edu/~scoulson/cogs179/kutas-hillyard80.pdf"
    },
    {
      "title": "Neurophysiological investigation of the basis of the fMRI signal",
      "url": "https://www.nature.com/articles/35084005"
    },
    {
      "title": "Inferring Causality from Noninvasive Brain Stimulation in Cognitive Neuroscience",
      "url": "https://pubmed.ncbi.nlm.nih.gov/32530381/"
    },
    {
      "title": "Causal Inferences in Repetitive Transcranial Magnetic Stimulation Research: Challenges and Perspectives",
      "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC7873895/"
    },
    {
      "title": "Cognitive Development in Childhood",
      "url": "https://nobaproject.com/modules/cognitive-development-in-childhood"
    },
    {
      "title": "How Children Discover New Strategies",
      "url": "https://www.routledge.com/How-Children-Discover-New-Strategies/Sieglerah2-Jenkinsah2/p/book/9781138972032"
    },
    {
      "title": "When does cognitive functioning peak? The asynchronous rise and fall of different cognitive abilities across the life span",
      "url": "https://pubmed.ncbi.nlm.nih.gov/25770099/"
    },
    {
      "title": "When does age-related cognitive decline begin?",
      "url": "https://pubmed.ncbi.nlm.nih.gov/19231028/"
    },
    {
      "title": "How learning to read changes the cortical networks for vision and language",
      "url": "https://pubmed.ncbi.nlm.nih.gov/21071632/"
    },
    {
      "title": "Learning to read alters cortico-subcortical cross-talk in the visual system of illiterates",
      "url": "https://pubmed.ncbi.nlm.nih.gov/28560333/"
    },
    {
      "title": "Guided participation in cultural activity by toddlers and caregivers",
      "url": "https://pubmed.ncbi.nlm.nih.gov/8284000/"
    },
    {
      "title": "Test-enhanced Learning",
      "url": "https://www.psychologicalscience.org/observer/test-enhanced-learning"
    },
    {
      "title": "The Effect of Type and Timing of Feedback on Learning From Multiple-Choice Tests",
      "url": "https://learninglab.psych.purdue.edu/downloads/2007/2007_Butler_Karpicke_Roediger_JEPA.pdf"
    },
    {
      "title": "Multi-Function Displays: A Guide for Human Factors Evaluation",
      "url": "https://www.faa.gov/data_research/research/med_humanfacs/oamtechreports/2010s/2013/201321"
    },
    {
      "title": "Human Factors Analysis of Safety Alerts in Air Traffic Control",
      "url": "https://hf.tc.faa.gov/publications/2007-human-factors-analysis/full_text.pdf"
    },
    {
      "title": "Reconstruction of Automobile Destruction: An Example of the Interaction Between Language and Memory",
      "url": "https://gwern.net/doc/psychology/cognitive-bias/1974-loftus.pdf"
    },
    {
      "title": "Identifying the Culprit: Assessing Eyewitness Identification",
      "url": "https://www.nationalacademies.org/projects/PGA-STL-13-02/publication/18891"
    },
    {
      "title": "When it Comes to Eyewitness ID Best Practices, the Science is Settled",
      "url": "https://cops.usdoj.gov/html/dispatch/12-2014/eyewitness_ID_best_practices.asp"
    },
    {
      "title": "The Replication Crisis in Psychology",
      "url": "https://nobaproject.com/modules/the-replication-crisis-in-psychology"
    },
    {
      "title": "PSYCHOLOGY. Estimating the reproducibility of psychological science",
      "url": "https://pubmed.ncbi.nlm.nih.gov/26315443/"
    },
    {
      "title": "Registered Reports",
      "url": "https://www.cos.io/initiatives/registered-reports"
    },
    {
      "title": "Preregistration",
      "url": "https://www.cos.io/initiatives/prereg"
    },
    {
      "title": "The weirdest people in the world?",
      "url": "https://doi.org/10.1017/S0140525X0999152X"
    },
    {
      "title": "The generalizability crisis",
      "url": "https://pubmed.ncbi.nlm.nih.gov/33342451/"
    },
    {
      "title": "How Computational Modeling Can Force Theory Building in Psychological Science",
      "url": "https://olivia.science/doc/guest_martin_21.pdf"
    }
  ],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}