RESEARCH DOSSIER
Individual Differences in Sensory Perception
Abstract. Individual differences in sensory perception concern systematic variation among observers in detection, discrimination, perceived intensity, multisensory combination, and interpretations of sensory experience. This fundamental research monograph synthesizes classical and modern psychophysics, measurement reliability, visual and auditory science, tactile function, molecular genetics of taste and smell, Bayesian cue integration, developmental change, and the evidence on sensory profiles in autism and attention-deficit/hyperactivity disorder. It distinguishes peripheral sensitivity, perceptual inference, decision criteria, affective response, self-reported environmental sensitivity, and functional behavior. Primary experimental studies, critical reviews, methodological research, and recent population analyses show how reliable perceptual variation may be specific to a sensory dimension, stimulus, or context. The synthesis identifies the reliability paradox, unisensory-baseline confounding, limitations of questionnaire-to-threshold inference, population sampling, and model identifiability as major methodological challenges. Future priorities include longitudinal, cross-cultural, and multimodal measurements, independent replication, and carefully defined comparisons with artificial perceptual systems.
Keywords: individual differences; sensory perception; psychophysics; perceptual sensitivity; sensory thresholds; visual perception; auditory perception; tactile perception; olfaction; taste genetics; multisensory integration; sensory processing sensitivity; autism; ADHD; measurement reliability; Artificial Era
Scientific Scope and the Problem of Observer Variation
Individual differences in sensory perception are reproducible variations among observers in detecting, discriminating, identifying, interpreting, and combining sensory information under specified conditions. They concern measurable performance, response criteria, and subjective experience across vision, hearing, touch, taste, smell, bodily sensation, and multisensory tasks. The scientific object is a distribution of person-specific perceptual functions rather than a ranking of persons by a single general sensitivity. A participant may discriminate color exceptionally well while showing ordinary temporal resolution or weak odor identification. Separating modality, task, and context is therefore essential. The methodological literature makes this heterogeneity a source of information about the architecture of perception Mollon et al. (2017).
An observation acquires the status of an individual difference when its between-person component survives reasonable scrutiny of within-person noise, stimulus variation, procedural artifacts, and sampling. A one-time unusually low threshold cannot establish an enduring sensory trait. Conversely, an effect that changes with lighting or attention may reveal a lawful person-by-condition interaction. The object of inquiry includes stable propensities, reversible states, experience-dependent tuning, biological variation, and their interaction. Psychophysical studies must specify what was physically delivered, what task was asked, and what statistical parameter was estimated; otherwise findings about apparent sensitivity may represent changes in criteria or reporting strategies Wichmann & Hill (2001a).
This monograph integrates historical psychophysics, sensory neuroscience, quantitative measurement, behavioral genetics, developmental studies, and phenomenological reports. The organizing question is how observers acquire distinct perceptual profiles despite sharing broadly similar human sensory systems. A second question concerns which differences generalize across time, stimulus classes, laboratories, and real environments. A third concerns whether established models predict differences at the level of individual parameters. The literature warrants a model of multiple interacting sources of variation, with explicit limits on causal inference. Current evidence about color appearance, audiovisual speech, receptor variants, and neurodevelopmental heterogeneity provides distinct tests of that claim Emery & Webster (2019) Basu Mallick et al. (2015).
From Classical Psychophysics to Differential Perceptual Science
Classical psychophysics established quantitative relations between stimulus properties and observers' judgments. Its enduring contribution was the possibility of estimating sensory functions with repeatable trials, graded intensities, discrimination comparisons, and specified decision procedures. The historical pursuit of a representative observer created standardized benchmarks while often treating differences across observers as error. Modern differential research instead asks whether those differences reflect systematic properties of receptors, neural encoding, learned categories, and decision policies. This methodological transition is visible in contemporary accounts of color science, which treat idiosyncratic settings as potential evidence about separate mechanisms rather than inconvenient deviations from a common mean Mollon et al. (2017) Emery & Webster (2019).
Threshold methods have changed because the result depends partly on how candidate stimuli are selected. Fixed stimulus levels offer transparent estimates but may be inefficient when individuals have widely separated performance functions. Staircases and adaptive Bayesian procedures concentrate measurements near informative intensity regions, improving efficiency under their assumptions. Leek's methodological review describes how adaptive trials use earlier responses to choose later stimuli and why different algorithms target different parameters. A threshold produced by an adaptive procedure remains an estimator with uncertainty, sensitivity to lapses, and possible sequential dependence. Comparison of two observers requires understanding the algorithm and the precision of both estimates Leek (2001).
Modern psychometric functions express response probability as a function of stimulus intensity or difference. Wichmann and Hill developed influential fitting, sampling, goodness-of-fit, and bootstrap approaches that support uncertainty-aware inference. Their work clarifies why apparent differences in a fitted threshold, slope, or lapse rate must be examined in light of sample size and model adequacy. Historical theories about fixed sensory limits can now be reexamined through hierarchical estimates and deliberate experimental manipulations. A coherent history of sensory individual differences therefore follows a progression from descriptive thresholds to statistical distributions, psychometric model testing, and mechanistic comparison Wichmann & Hill (2001a) Wichmann & Hill (2001b).
Constructs, Levels of Explanation, and Scientific Boundaries
Sensory detection refers to the probability of registering a stimulus against a background; discrimination concerns distinguishing stimuli; identification attaches a category or source label; intensity judgments place experience on a response scale; and sensory affect concerns pleasantness, aversion, or salience. These outcomes can diverge within one person. An observer who reliably detects an odor may dislike it intensely, while another may assign the same intensity but a different qualitative identity. Sensory-processing sensitivity as assessed by personality questionnaires samples broader environmental and emotional responsiveness, and its scores cannot be read as direct physical detection thresholds Aron & Aron (1997) Greven et al. (2019).
Analysis benefits from distinguishing peripheral transduction, neural representation, perceptual inference, response selection, and reports about everyday life. Genetic differences in a receptor can influence input transduction, whereas expectations can influence inferred source identity or the interpretation of ambiguous evidence. Attentional capacity, fatigue, learning history, and language affect subsequent operations. The resulting perceptual profile is therefore multidimensional and may have several causal pathways. Direct psychophysics, standardized questionnaires, physiological recording, and interviews each describe different parts of this organization. Evidence from experimentally manipulated visual–haptic reliability demonstrates that perceptual weighting is conditional on signal quality rather than permanently assigned to a modality Ernst & Banks (2002).
The boundary with neighboring research areas should remain explicit. Sensory thresholds and discrimination focus on foundational detection functions; specific hearing or visual disorders require their own diagnostic frameworks; sensory adaptation examines changes produced by recent exposure; and personality science investigates broad dispositional patterns. This dossier examines covariation among those phenomena at the level of the person. Clinical status is one candidate explanatory variable, not the organizing definition of perceptual difference. A population-based comparison must preserve variation within each diagnostic group, and a group mean cannot establish a sensory phenotype shared by every member Robertson & Baron-Cohen (2017).
Sources of Variation: Persons, Trials, Stimuli, and Contexts
A recorded response arises jointly from person, stimulus, moment, measurement apparatus, and response instructions. Between-person variance is the component of interest only when the question targets stable individual differences. Within-person variation includes spontaneous fluctuations, changes in attention, fatigue, adaptation, and error arising from limited observations. Stimulus variation may be especially consequential when nominally identical categories contain exemplars with distinct acoustic, spatial, or chemical structure. For the McGurk effect, Mallick and colleagues measured dramatic differences across people and video exemplars, finding that averages could conceal nearly categorical differences in response across the stimulus set Basu Mallick et al. (2015).
A useful repeated-measures study deliberately crosses observers with several stimuli and multiple sessions. Such a design separates person effects, item effects, session effects, and person-by-item interactions, although complete causal identification still depends on randomization and the experimental context. Using just one highly effective illusion exemplar creates a narrow definition of sensitivity. Using many exemplars without accounting for their differences can obscure person-specific tuning. Strand and colleagues found that audiovisual illusion susceptibility was related to some measures of lipreading and incongruity detection, reinforcing the need to assess the constituent abilities before assigning a global integration score Strand et al. (2014).
Perceptual context encompasses adaptation level, environmental illumination, background noise, stimulus history, semantic expectations, and task demands. These factors can be manipulated to distinguish robust observer parameters from conditional responses. For color vision, optical factors and photoreceptor variability do not map straightforwardly onto color appearance because visual systems partly compensate for their inputs. Between-person variation can consequently be large in one hue range and modest in another. A serious individual-difference analysis must report the sampling range across stimuli, sessions, and contexts, allowing the observer effect to be interpreted at its actual level of generality Emery & Webster (2019).
Measurement Reliability and the Reliability Paradox
Reliability is the reproducibility of a score for the inference being made. Test–retest correlations address stability across occasions; split-half estimates concern consistency across items; and hierarchical variance components locate uncertainty in observers, trials, and stimuli. High accuracy or a robust average experimental effect does not guarantee a reliable measure of individual differences. Hedge, Powell, and Sumner demonstrated that familiar experimental tasks can show robust group effects while providing weak rank-order reliability, partly because participants resemble one another on the measured difference. This reliability paradox changes how cognitive and sensory scores should be selected for correlational explanations Hedge et al. (2018).
Before correlating one sensory threshold with another trait, researchers should inspect the threshold's distribution, trial count, ceiling and floor behavior, and standard error. Two unreliable measures can produce misleadingly weak observed associations; correcting an observed correlation for reliability requires additional assumptions and reliable estimation of both measurement errors. If the participant score is an average over heterogeneous stimuli, an aggregate reliability coefficient may conceal inconsistency across stimulus classes. Study preregistration should specify the intended estimand and generalization universe. Independent replication must repeat the target measurement, rather than merely obtaining the same average condition effect Mollon et al. (2017) Wichmann & Hill (2001b).
Reliability can coexist with narrow scope. A participant may respond consistently to a specific audiovisual syllable pair without showing equivalent responses to a different speaker, while another parameter remains stable across stimuli. Basu Mallick and colleagues reported stability over a one-year retest despite large between-person and between-stimulus variability, illustrating why reliability and universal task invariance are distinct questions. Precision can be improved through more trials, multiple calibrated items, and model-based estimation, yet no numerical transformation can rescue a construct that mixes unrelated processes without justification. Validation remains a substantive theory-dependent task Basu Mallick et al. (2015) Hedge et al. (2018).
Visual Sensitivity and Color Appearance
Visual individual differences emerge before cortical interpretation: ocular optics, lens coloration, macular pigmentation, cone photopigments, and retinal sampling vary across observers. These sources change the effective spectral signal reaching visual mechanisms even when external illumination is held constant. Yet color experience cannot be inferred directly from receptor sensitivity. Opponent processes, adaptation, normalization, and learned reference categories influence perceptual equality settings. Emery and Webster describe both the diversity of color judgments and compensatory processes that stabilize appearance. This combination explains why equal discrimination thresholds do not guarantee equal hue experiences and why spectral sensitivity is an incomplete explanation of perceptual phenotype Emery & Webster (2019).
Psychophysical color tasks measure different constructs. Color matching constrains relations between physical spectra and response equivalence; discrimination tasks estimate minimal detectable differences; hue scaling samples qualitative appearance; and naming introduces language-dependent category boundaries. Individual hue-scaling functions can vary across restricted color regions rather than as one uniform offset. For a meaningful comparison, researchers need calibrated displays, controlled luminance and adaptation, validated observer instructions, and explicit consideration of anomalous color vision. The visual-science methodological review stresses that correlations among tests can illuminate separate mechanisms but become difficult to interpret when task artifacts or stimulus ranges differ Mollon et al. (2017).
Stable differences in color experience raise a central theoretical issue: what is held constant when the same named color is reported by different observers? Agreement on labels may follow shared linguistic training while perceptual balance points differ. Disagreement in appearance can occur without reduced visual acuity. Such results require separating sensory physiology from the category systems used in reporting, and distinguishing interobserver comparison from change within an observer after adaptation. Applied consequences arise in color-critical design, scientific visualization, industrial inspection, and accessible communication. Experimental reports should publish physical stimulus specifications and avoid treating the average normal observer as a complete description of the population Emery & Webster (2019).
Auditory Resolution, Pitch, and Speech in Noise
Audition contains several partially distinct abilities: detecting faint sounds, discriminating frequency and temporal changes, localizing sources, recognizing pitch patterns, and recovering speech in competing sound. These functions draw on cochlear filtering, temporal coding, central grouping, attention, and experience. Oxenham's synthesis of auditory coding explains why cochlear frequency selectivity is crucial to pitch and auditory scene analysis. It also clarifies why a single pure-tone audiogram cannot fully characterize an individual's performance with music, reverberation, or multiple talkers. The scientific question concerns the distribution of performance across acoustically specified tasks, particularly when ordinary hearing thresholds coexist with difficulty in complex listening conditions Oxenham (2018).
Auditory comparisons require calibrated sound pressure, headphone or loudspeaker characteristics, room acoustics, frequency range, hearing history, and age. Performance on pitch labeling is influenced by musical learning and the structure of response categories; fine pitch discrimination can be studied without requiring note names. Temporal order and gap detection involve further parameters and may show different relation to speech perception. Even when two participants produce a similar aggregate score, distinct patterns of error may reveal differences in spectral resolution, temporal analysis, or attention. Adaptive estimation supports efficient individual measurement, but its convergence behavior must be compared with independent fixed-level trials when distributions are unusually broad Leek (2001) Oxenham (2018).
Listening in everyday settings adds cognitive load, familiarity with speakers, and prediction about the topic. An auditory advantage under one listening condition may disappear when signals become unfamiliar or speech and background noise change. Research programs should therefore report several listening environments rather than equating performance with a permanent general auditory aptitude. Crossmodal cues can partly compensate for uncertainty in sound localization or speech identification, and their use is itself variable. Treating unisensory auditory precision as a separate parameter is necessary before drawing conclusions from the size of any audiovisual perceptual effect Alais & Burr (2004) Strand et al. (2014).
Tactile Acuity, Body Sensation, and Somatosensory Profiles
Somatosensation combines cutaneous pressure, vibration, texture, temperature, and nociceptive signaling with proprioceptive information about limb position and movement. Individual profiles can differ across body sites, stimulus intensities, and tasks. Spatial discrimination on a fingertip measures a different system configuration from judging a sustained vibration on the forearm or feeling pressure on the skin. Receptor density, peripheral conduction, central receptive fields, exposure to skilled activity, and attention are plausible contributors. The observation of different thresholds should therefore be interpreted through the tested site and stimulus rather than a global classification of tactile sensitivity Ernst & Banks (2002) Mollon et al. (2017).
Touch often participates in action rather than passive report. A person handling an object samples texture and shape through self-generated movements; perceptual precision depends partly on exploration strategy and motor control. Ernst and Banks measured separate visual and haptic precision before comparing their weighted combination, thereby creating a method for disentangling differences in tactile input from integration. This distinction also matters in proprioceptive tasks: a poorly calibrated joint-position estimate can alter sensorimotor performance even when the person describes touch sensation as normal. Functional measurement should use comparable movement opportunities, calibrated contact forces, and repeat measurements across sessions Ernst & Banks (2002).
Subjective tactile distress and laboratory tactile thresholds need not coincide. Everyday experiences involve unpredictability, social meaning, texture preferences, and anticipatory responses, any of which may influence avoidance without showing altered elementary detection. Questionnaire profiles provide important ecological observations when interpreted as reports of daily behavior; they cannot alone identify receptor or cortical mechanisms. Developmental norms and repeated direct testing are needed when evaluating individual change. The classic Sensory Profile research established structured sampling of childhood sensory responses while illustrating the separate question of how representative norms transfer to particular contexts and populations Dunn & Westman (1997).
Taste Genetics, Bitterness, and Chemical Context
Human taste variation offers unusually clear examples of biological mechanisms linking receptor differences to psychophysical outcomes. The TAS2R38 bitter receptor system is associated with differences in perceiving phenylthiocarbamide and related compounds. Kim and colleagues mapped a human quantitative trait locus for PTC sensitivity, and Bufe and colleagues combined receptor function with participant responses to establish molecular pathways for specific bitter compounds. These findings demonstrate that a genetic variant can influence a perceptual threshold under defined chemical conditions. They do not support an inference that a person who tastes PTC strongly must be unusually sensitive to every flavor, nutrient, or environmental stimulus Kim et al. (2003) Bufe et al. (2005).
Chemical stimulus identity is central to construct validity. Different bitter molecules activate partly different receptor repertoires, and concentration, solvent, temperature, saliva, recent food intake, and adaptation alter the delivered stimulus. Threshold detection and suprathreshold intensity ratings can produce different individual rankings. A person may distinguish minute concentrations but use a conservative intensity scale, while another uses larger numerical ratings without better detection. Reliable characterization therefore combines controlled dilution series, forced-choice discrimination, repeated sessions, and explicit evaluation of response scaling. Studies linking genotype with taste should report ancestry, allelic distribution, and environmental covariates because receptor effects have domain-specific limits Bufe et al. (2005).
Food choice extends the analysis beyond the receptor. Learned preferences, cultural norms, accessibility, food processing, and health concerns can alter consumption even when bitterness is strong. A genetically informative taste test is not itself a causal explanation of dietary patterns. The broader research tradition distinguishes molecular receptor function, psychophysical perception, affective evaluation, and actual behavior. Human differences in sensitivity to bitter foods are particularly instructive because a relatively direct receptor effect exists alongside contextual influences on preference. Integrative experiments must measure each intermediate variable before treating any reported eating preference as the expression of sensory biology Kim et al. (2003) Bufe et al. (2005).
Olfactory Receptors, Odor Quality, and Learning
Olfaction illustrates the combinatorial character of perception: odorant molecules activate overlapping receptor repertoires and the resulting patterns are interpreted through neural processes. Keller and colleagues demonstrated that genetic variation in a human odorant receptor alters perception of particular odorants, providing a mechanistic connection between molecular variation and reported odor quality. Their work with OR7D4 and androstenone shows why observers may differ radically in the intensity or character of the same odor. Such a finding concerns a defined receptor–odorant relation; it cannot be generalized into a single genetically fixed ability to smell well across the entire odor landscape Keller et al. (2007).
Experimental odor perception is vulnerable to contamination and adaptation. Concentration delivery, airflow, sniff magnitude, exposure duration, odorant purity, residual cues, and repeated presentation all matter. Identification depends further on language, prior exposure, and familiarity, whereas detection can be tested with unfamiliar molecules. This creates a useful dissociation: somebody may detect an odor correctly yet fail to name it, or identify a familiar scent from contextual cues despite weak sensitivity in a new task. A complete sensory profile would analyze detection, intensity, quality, hedonic valence, and naming as distinct outcomes and report repeat reliability within each outcome Keller et al. (2007) Mollon et al. (2017).
Everyday olfactory environments contain mixtures, expectations, and emotion-laden associations. Experiences of fragrance, cooking, and environmental threat may change through repeated exposure, illness, or ordinary variation in nasal conditions. Genetic effects coexist with training and contextual interpretation; a receptor difference may shape one aspect of input while memory controls the label and affective meaning. Cross-modal food flavor adds gustatory and trigeminal components. Experiments designed to explain individual odor judgments should specify whether the target variable is chemical detection, perceptual categorization, affective response, or functional use in natural environments. This precision prevents turning a specific psychophysical finding into an unqualified claim about personal sensitivity Keller et al. (2007).
Signal Detection, Criteria, and Perceptual Decisions
Signal detection analysis separates the quality of evidence available for a decision from the criterion an observer uses to declare that a stimulus was present. Two individuals can respond differently to identical weak sounds because one requires stronger evidence before saying yes, rather than because their auditory encoding differs. Hit and false-alarm rates jointly contribute to sensitivity and criterion measures under specified distributional assumptions. Unequal-variance models and alternatives become relevant when data violate a simple Gaussian formulation. Person comparisons based only on raw detection percentages conceal such distinctions. The scientific estimand must specify discrimination, decision bias, or both Mollon et al. (2017) Wichmann & Hill (2001a).
Forced-choice procedures restrict some response biases, but do not remove attention, strategic learning, memory limitations, or item-specific preferences. In a two-alternative task, performance reflects both available evidence and the decision process. If an observer lapses occasionally, psychometric slope may become misleading unless lapses are estimated separately. Wichmann and Hill advocate fitting functions with attention to goodness of fit, sampling, and uncertainty; treating a fitted threshold as a direct neural measurement discards the measurement model. Confidence judgments can enrich a decision analysis, yet confidence calibration adds another task and should be validated as a separate person-level characteristic Wichmann & Hill (2001a) Wichmann & Hill (2001b).
Within-participant manipulations provide leverage for distinguishing sensory and decisional accounts. Changing stimulus signal-to-noise ratio tests evidence quality, while changing payoff or stimulus probability probes criterion flexibility. Cross-session retesting examines whether the derived parameters are stable enough to support individual comparisons. A participant who shifts criterion predictably when incentives change may exhibit intact sensory precision even if baseline yes responses are uncommon. This explanation should be preferred over attributing all response frequency differences to sensory deficit. Individual-difference research benefits from reporting full response distributions and the fitted parameters that connect them to a specified generative account Hedge et al. (2018).
Reliability-Weighted Multisensory Cue Combination
Vision, hearing, and touch provide partly redundant estimates of environmental properties. Reliability-weighted integration predicts that an estimate based on a more precise cue should receive greater weight, assuming an appropriate common source and statistical structure. Ernst and Banks measured visual and haptic precision separately and found combination behavior resembling a maximum-likelihood estimator. Alais and Burr observed that changing the blur of a visual cue altered its relative influence over auditory localization, moving from visual dominance to greater auditory influence. These studies demonstrate a conditional computational principle and explain why differences in baseline acuity can produce differences in apparent cue dominance Ernst & Banks (2002) Alais & Burr (2004).
For two independent unbiased Gaussian cues, a simple ideal observer combines estimates with weights inversely proportional to their variances; the predicted combined variance is lower than either isolated variance. This relationship provides a benchmark, not a guarantee that human behavior is optimal under every condition. Correlated errors, unequal costs, incorrect causal assumptions, decision noise, and learned biases all alter performance. Quantifying each observer requires estimating unisensory variance before interpreting the combined response. Apparent superiority on a multisensory task may reflect exceptionally good vision or touch rather than an independently superior integration mechanism Ernst & Banks (2002) Zhu et al. (2026).
Development adds another qualification. Gori and colleagues found that children under approximately eight years did not combine visual and haptic form information in the same statistically optimal fashion observed later; performance depended strongly on task and dominant modality. This study concerns specific spatial discrimination paradigms and illustrates that integration policies develop. A developmental individual difference can therefore reflect differences in maturation, recalibration history, or available sensory precision. Longitudinal sampling is needed to distinguish a persistent characteristic from a transient stage. Researchers should avoid assigning a lifelong trait on the basis of a single age-specific measurement Gori et al. (2008).
Bayesian Causal Inference and Perceptual Priors
Reliability-weighted averaging assumes signals are relevant to a common event. Bayesian causal inference extends the problem by estimating whether two observations arise from one source or from separate sources before choosing a perceptual interpretation. Observers may differ in sensory noise, learned expectations about co-occurrence, assumptions about temporal coincidence, or decision rules. A high rate of audiovisual fusion can therefore arise through several parameter combinations. The framework offers a conceptual explanation for heterogeneity while demanding experimentally identified parameters. Zhu, Beierholm, and Shams have emphasized that the raw magnitude of a multisensory illusion cannot by itself identify the integration mechanism Zhu et al. (2026).
Perceptual priors are expectations about states of the world that influence inference when sensory evidence is uncertain. Their effect depends on the likelihood structure and the decision task. If cues are nearly congruent, many parameter settings can predict similar responses; incongruent cues across graded levels of reliability are more diagnostic. An observer who uses stronger prior expectations may show more decisive perception under ambiguity, yet this need not indicate stronger peripheral sensitivity. A scientific comparison should include conditions in which the evidence supports both common-source and separate-source interpretations, and should test whether model parameters generalize across contexts Zhu et al. (2026) Alais & Burr (2004).
Individual differences in priors are difficult to estimate independently of perceptual precision when stimuli are limited. Different generative models can fit the same response proportions, and high flexibility may obscure identification. Model comparison, out-of-sample prediction, parameter-recovery simulations, and independent baseline conditions improve interpretability. Hierarchical modeling allows partial pooling while retaining uncertainty about each participant. The central theoretical lesson is that perceptual differences emerge from an inference system with multiple components. Researchers gain explanatory power by testing these components rather than attributing a visible difference immediately to generalized hypersensitivity, weak integration, or a single cognitive style Zhu et al. (2026).
Sensory Adaptation, Calibration, and Perceptual Learning
Sensory systems change their responses as recent stimulation changes. Visual adaptation can recalibrate sensitivity and perceived neutrality; repeated sound exposure can alter expectations; and sustained contact can transform tactile experience. Such effects complicate the distinction between stable traits and momentary states. A person tested after exposure to bright illumination may perform differently from the same person tested after darkness, without any change in enduring visual constitution. Experimental control should therefore include adaptation duration, intertrial intervals, background statistics, and history-dependent effects. The interaction between physically variable retinal inputs and compensatory visual mechanisms offers one illustration of how recalibration stabilizes color experience Emery & Webster (2019).
Perceptual learning changes discrimination or identification performance with experience. Repetition may improve a practiced task by refining representations, attention to diagnostic features, or decision policies, and transfer to other tasks can be limited. Apparent between-person aptitude may partly index unequal prior learning, especially when expert listeners, visual specialists, or trained tasters are compared with novices. A meaningful design records expertise and exposure history, trains observers to comparable asymptotes when appropriate, and reports learning trajectories rather than only post-training ranks. This approach makes it possible to ask whether differences persist after task familiarity is equated Mollon et al. (2017).
Adaptation and learning have different temporal implications, yet both operate within the same observer. Multi-session experiments can manipulate exposure and measure recovery, while longitudinal experiments can track durable change after training. The resulting models should include participant-specific baselines and response functions over time. If an individual's apparent sensitivity changes selectively for one stimulus family, general claims about a global sensory trait are weakened. The observed plasticity itself becomes a legitimate individual-difference phenotype: persons may differ in the rate, size, and specificity of adjustment. Distinguishing those parameters requires dense time sampling and reproducible experimental procedures Leek (2001) Emery & Webster (2019).
Attention, Expectation, and the Interpretation of Experience
Detection and discrimination depend on where attention is allocated, which alternatives are anticipated, and whether stimulus features are relevant to the task. Perceptual sensitivity measured during divided attention can differ from sensitivity measured with an explicit location cue. This variation may reflect selective resource allocation, strategic response control, or changes in the effective evidence available for a judgment. Researchers should assess attention independently rather than interpret every between-person threshold difference as a difference in primary sensation. Comparing multiple tasks allows estimation of the shared and task-specific components, provided the measures have adequate reliability Hedge et al. (2018).
Expectation also changes what an ambiguous sensory event is taken to mean. A person anticipating speech may perceive a noisy sound as meaningful; familiar odor labels can guide recognition; social context can change the reported unpleasantness of touch. A subjective report is scientifically important but combines the sensory event with its interpretation and response vocabulary. Objective discrimination tasks and controlled expectancy manipulations help determine where an effect enters the processing chain. Experiments involving audiovisual speech show how responses to identical physical signals can depend on task format and the observer's lipreading capacity, even when sensory inputs themselves remain fixed Strand et al. (2014) Basu Mallick et al. (2015).
Attention and affect may also influence sensory-related avoidance or seeking. Repeated exposure to environments perceived as difficult can change anticipatory behavior, while deliberate active exploration can change available information. A questionnaire item about avoiding crowded rooms therefore contains contextual, emotional, and social information, not just a pure-tone or light-detection threshold. Greven and colleagues situate sensory-processing sensitivity within a broader environmental-sensitivity research program. Integrating that program with psychophysics requires explicit tests of convergent and discriminant validity, including whether reported experience predicts sensory performance after attention, affect, and environmental exposure are considered Greven et al. (2019) Aron & Aron (1997).
Developmental Trajectories and Lifespan Heterogeneity
Individual differences appear in childhood within a changing perceptual system. Maturation of sensory pathways, motor exploration, attentional control, and learned categories can change the meaning of the same test score across ages. Developmental measurement must distinguish chronological age, developmental opportunity, task comprehension, and sensory function. Gori and colleagues' visual–haptic work shows that children can favor one modality in particular form tasks before later developing reliability-weighted combination patterns. Their results support a developmental account of integration while leaving open the precise longitudinal pathways within individuals. Cross-sectional age differences should not be interpreted as direct observations of personal change Gori et al. (2008).
Childhood sensory questionnaires capture participation in home, school, and social contexts. Dunn and Westman's normative study of the Sensory Profile demonstrated the feasibility of systematic descriptions in a general child sample, but such reports reflect how environments and caregivers interact with a child's responses. Physical thresholds require additional controlled tests. Individual profiles can also be shaped by educational access, noise exposure, developmental experience, and learning. A child showing distress in a noisy classroom might have atypical auditory encoding, unusually high attentional load, a salient previous experience, or a combination. Longitudinal multi-method assessment is needed to identify these explanatory components Dunn & Westman (1997) Greven et al. (2019).
Later-life changes in optics, cochlear function, proprioception, and processing speed influence sensory measurements, often at different rates and under different conditions. A biological change that affects one person's hearing may increase their reliance on vision for audiovisual comprehension; another observer may adapt through practice or environmental modification. Age distributions should be modeled rather than treated as nuisance covariates. Cohort effects and selective participation further complicate comparison of age groups. Study design should ideally sample repeated perceptual functions over years, record relevant medical histories, and distinguish general age trends from substantial variation among individuals of the same age Oxenham (2018) Emery & Webster (2019).
Neurodevelopmental Variation and Autism
Autistic sensory experience is heterogeneous across modalities and contexts. Reports can include distress from intense environments, reduced reaction to some cues, strong discrimination in selected tasks, or differences in social stimulus processing. Robertson and Baron-Cohen's review places these phenomena within sensory and neural research rather than reducing them to one general direction of sensitivity. Empirical studies vary in their stimuli, response formats, participant characteristics, and cognitive demands; a group contrast therefore needs a well-specified target outcome. Any model of autistic sensory perception must allow individual profiles to differ from the group mean and from one another Robertson & Baron-Cohen (2017).
Hadad and colleagues combined psychophysical observations from 107 autistic and 408 comparison participants across 32 conditions that included brightness, orientation, pitch, size, faces, and speech. Their 2026 analysis used segmented regression and Bayesian hierarchical modeling and reported a domain-specific pattern: average performance in face and speech tasks differed from that of the comparison group, while basic nonsocial conditions were comparable or sometimes favored autistic participants. This result illustrates the value of modeling heterogeneous domains and participants. Its pooled, task-specific findings do not establish an invariant sensory signature or explain every participant's daily experiences Hadad et al. (2026).
Questionnaire findings and controlled psychophysics address different levels. Sensory avoidance may be strongly related to predictability and environmental consequences even when thresholds are within a comparison distribution. Conversely, a reliable laboratory advantage may not be experienced as helpful in a noisy public setting. Reports should distinguish threshold, intensity, perceptual organization, attentional filtering, distress, and functional impact. Samples should include diverse communication abilities and accommodate participants' response needs. The most informative scientific question is which measured mechanisms account for which individual outcomes, rather than whether an entire diagnostic group is globally more or less sensitive Hadad et al. (2026) Robertson & Baron-Cohen (2017).
ADHD and the Interpretation of Sensory Questionnaire Evidence
Studies of attention-deficit/hyperactivity disorder frequently examine sensory seeking, avoidance, registration, and sensitivity through self-report or observer-report instruments. These dimensions describe experiences and behavior in daily environments; their relation to laboratory discrimination functions requires independent testing. Jurek and colleagues' 2025 systematic review and meta-analysis assembled 30 studies with 5,374 participants and reported higher questionnaire-based sensory atypicalities in ADHD groups than comparison groups across several categories. The synthesis provides population-level evidence of association, while the included studies' methods and heterogeneity constrain mechanistic conclusions. Questionnaire scores alone do not demonstrate peripheral sensory hyperacuity Jurek et al. (2025).
Attention is particularly relevant because it changes whether a sensory event is noticed, remembered, or judged important. Failure to report a weak signal can reflect inattention during a trial, low encoding precision, conservative decision criteria, or competing environmental demands. Sensory seeking may likewise reflect reward, self-regulation, situational boredom, or sensation preference. To separate these possibilities, research should compare direct psychophysical estimates with attentional performance, repeated observer reports, and daily-context measures. The overall association reported in a meta-analysis is a starting point for constructing discriminating hypotheses rather than a complete causal account Jurek et al. (2025) Hedge et al. (2018).
Group comparisons should also consider medication status when relevant, co-occurring conditions, age, sleep, stimulus familiarity, and reporting context. Such variables may influence sensory judgments or daily-life experiences and can differ systematically across recruited samples. Measurement invariance is essential: a questionnaire item must have sufficiently similar meaning across populations before its raw score supports a precise group comparison. Accessible environments can be studied experimentally by modifying noise, illumination, pace, or opportunities for voluntary exposure. The preferred endpoint is a specified improvement in functioning or comfort measured directly, rather than an assumed correction of an unmeasured sensory mechanism Jurek et al. (2025).
Sensory-Processing Sensitivity and Personality Measurement
Sensory-processing sensitivity is a personality-oriented construct describing differential responsiveness to environmental influences, including emotional and contextual features. Aron and Aron introduced a measurement program in 1997 that sought to distinguish this trait from introversion and emotionality. Its conceptual scope spans more than low-level visual, auditory, tactile, gustatory, or olfactory thresholds. Greven and colleagues' critical review situated it within environmental sensitivity and highlighted both the promise of trait research and the need for stronger objective measurement. A high score on a sensitivity questionnaire describes a tested self-report construct; it should not be presented as a clinical diagnosis or universal indicator of sensory superiority Aron & Aron (1997) Greven et al. (2019).
The current measurement debate concerns dimensionality and construct overlap. Early work emphasized a broad underlying factor, while later research considered distinguishable components involving subtle cues, overstimulation, depth of processing, and emotional responsiveness. In 2026, Pluess and colleagues introduced a revised Highly Sensitive Person scale using a U.S.–U.K. adult sample of 1,000 and a six-dimension structure. The instrument's validation provides evidence about questionnaire factors and their relationships to other traits, while independent replication and cultural invariance remain consequential requirements. This development illustrates why a questionnaire label cannot be assumed to have a constant meaning across successive versions Pluess et al. (2026).
Linking personality measures to sensory performance requires studies that administer both kinds of instruments within the same participants. Researchers should specify whether they expect low detection thresholds, higher rated intensity, greater perceptual awareness, stronger affect, or greater susceptibility to environmental stress. These are testably different outcomes. The reliability of the physical measures and discriminant validity of personality questionnaires jointly constrain any observed association. Broad trait measures may be useful for understanding environmental accommodation or wellbeing even when they do not predict elementary sensory discrimination. The critical goal is a calibrated explanatory connection between reports, functional behavior, and sensory operations Greven et al. (2019) Hedge et al. (2018).
Genes, Environment, and the Limits of Heritability
Behavioral genetic analysis of sensory differences should identify the phenotype before estimating genetic influence. Molecular studies link variants at TAS2R38 to responses to particular bitter compounds, and variants of the OR7D4 odorant receptor to perception of specified odors. These are concrete mechanistic examples that can support causal pathways under appropriate experimental conditions. They also reveal the narrowness of what is explained: a receptor genotype cannot account for the full cognitive, social, and environmental organization of a person's sensory experience. Gene effects may interact with exposure, response context, and the receptor repertoire for each stimulus Bufe et al. (2005) Keller et al. (2007).
Heritability is a statistical description of variation attributable to genetic differences in a specified population and environment. It is not the percentage of an individual's perception caused by genes, nor a claim that a trait cannot change. A highly heritable sensory phenotype can remain modifiable through disease, learning, or environmental conditions. Genetic associations may differ across ancestry groups when the distributions of variants and environments differ. Phenotype measurement error can further obscure true association, and candidate-gene narratives require independent replication and appropriate control of multiple testing. The broader differential literature advocates treating biological variation as one explanatory layer among others Mollon et al. (2017).
Population sampling is especially important when genomic results are translated into claims about human sensory diversity. Convenience samples from one geographic area may underestimate rare receptor variants and oversimplify cross-population distributions. Even well-established receptor effects can fail to predict complex food choices because perceptions coexist with learning and social practices. Strong inference triangulates genotype, receptor function, controlled perception, and naturalistic behavior rather than skipping intermediate levels. The research opportunity is to build mechanistically specific profiles that explain when biological differences become observable and when environmental compensation or learning reduces their behavioral expression Kim et al. (2003) Keller et al. (2007).
Population Validity, Cultural Context, and Measurement Invariance
An individual-difference estimate generalizes only to populations and environments represented in the research design. Studies recruiting university volunteers may provide precise data for selected sensory paradigms while omitting age ranges, hearing histories, occupational exposures, disability experiences, and linguistic backgrounds. Measurement invariance asks whether a task or questionnaire expresses sufficiently comparable constructs across groups. The same verbal category can capture different sensory or emotional meanings when language and experience differ. Research on color perception particularly demonstrates the need to distinguish physical matching, perceptual appearance, and linguistic labels before comparing group distributions Emery & Webster (2019) Mollon et al. (2017).
Sampling should make explicit who can complete the task. Hearing- or vision-dependent instructions can inadvertently exclude the very participants whose perceptual differences are scientifically central. Translation of questionnaire items requires more than literal word substitution, and statistical factor equivalence cannot automatically be assumed across languages. Accessibility adjustments should be recorded as part of the measurement procedure, with their effect on comparability evaluated rather than ignored. Experimental reports should characterize the chosen population and its exclusions so readers can assess transportability. Traits inferred from narrow samples should remain tied to those samples until independently replicated in new settings Greven et al. (2019) Pluess et al. (2026).
Population means can also conceal mixtures and extreme profiles. A subgroup may show reliable enhancement on one stimulus class while another shows reduced accuracy, leaving the aggregate close to zero. Hadad and colleagues' domain-sensitive analysis of autistic and comparison participants illustrates why heterogeneous modalities require more than a global group contrast. Multilevel models can estimate participants and conditions simultaneously, but their conclusions still depend on sampling and model adequacy. Replication should include novel exemplars and independent laboratories. The appropriate scientific ambition is a transferable explanation of variation, supported by measurement equivalence and external validation, rather than an attractive population stereotype Hadad et al. (2026).
Experimental Programs, Statistical Models, and Research Transparency
A rigorous program starts with a narrow construct, a physical stimulus specification, and a parameter that can answer the hypothesis. Threshold studies should collect sufficient near-threshold observations to estimate a psychometric function with uncertainty. Multisensory studies need unisensory baselines and graded cue conflict. Questionnaire studies need validation and comparison against independent outcomes. Genetic studies require replicable genotype–phenotype associations and appropriate population controls. Model choices should follow the scientific question, with trial-level data retained when possible. Wichmann and Hill's fitting and confidence-interval analyses provide a methodological foundation for estimating uncertain perceptual parameters instead of treating point estimates as observations free of error Wichmann & Hill (2001a) Wichmann & Hill (2001b).
Hierarchical models estimate distributions of observer parameters while accounting for within-observer uncertainty and partial pooling. Crossed stimulus effects separate person variation from item-specific elicitation. A preregistered reliability study should repeat measurements under conditions matched to the desired use and, when feasible, use an independent stimulus set. Model comparison should incorporate parameter recovery and predictive validity, particularly where priors, likelihood precision, and response strategies can mimic one another. The reliability paradox described by Hedge and colleagues warns against assuming a robust experimental mean supplies a useful stable score for correlational research. Investigators should report negative findings and uncertainty about trait-level inferences Hedge et al. (2018) Zhu et al. (2026).
Evidence synthesis must preserve distinctions between study designs. Controlled laboratory experiments can support causal claims about manipulated stimulus properties within their conditions; correlational questionnaire studies describe associations; family and genetic studies illuminate particular biological pathways; systematic reviews summarize heterogeneous results subject to inclusion and risk-of-bias constraints. A narrative monograph integrates such materials critically but does not acquire the status of a systematic review merely by citing many publications. Clear provenance, author–year citations, stable DOI links, and an explicit account of limitations make synthesis auditable. The governing question remains what was measured and what inferential step is justified by each study Jurek et al. (2025) Greven et al. (2019).
Everyday Function, Sensory Accessibility, and Accommodation Research
Laboratory functions matter when they help explain participation in education, work, travel, communication, leisure, and public environments. Accessibility interventions can modify acoustic intensity, lighting, tactile features, visual contrast, or the predictability and timing of signals. Evaluation should identify the intended functional outcome and measure whether a specific adjustment improves performance, safety, comfort, or choice. A universal setting is rarely optimal across all observers because sensory capabilities and preferences vary. An individual's direct report is essential for selecting practical outcomes, while calibrated tests can clarify which physical changes are relevant Dunn & Westman (1997) Mollon et al. (2017).
Audio–visual communication provides an instructive case. Lipreading can assist speech understanding, but people differ in how they combine congruent and discrepant visual cues with sound. Strand and colleagues related illusion susceptibility to individual differences in lipreading and incongruity detection, while Mallick and colleagues showed stable yet highly stimulus-specific variability. Accessibility research should therefore test captions, visual cues, audio clarity, signal redundancy, and user control on their actual communication outcomes. The size of one audiovisual illusion is not by itself a measure of how useful an accommodation will be in conversation Strand et al. (2014) Basu Mallick et al. (2015).
Ethical interpretation requires avoiding unnecessary pathologization of human variation while recognizing that sensory environments can impose genuine functional limitations. A strong reaction to unexpected sound or touch is an important phenomenon regardless of whether elementary laboratory thresholds differ. Any clinical assessment depends on the complete health and functional context, not one sensory score. Population-level evidence about autism and ADHD should guide testable hypotheses and inclusive design, while avoiding automatic assumptions about a particular person's needs. Scientific and practical quality improve together when environments are adjustable, outcomes are documented, and observers participate in interpreting their own experiences Robertson & Baron-Cohen (2017) Jurek et al. (2025).
Comparative Perception and the Artificial Era
Artificial perceptual systems make individual-difference questions experimentally tractable in a second technological domain. Distinct sensors, input preprocessing, model architectures, learned representations, and calibration histories can yield different outputs to identical external scenes. The meaningful comparison with human perception occurs at the level of operational tasks: discrimination under noise, response stability, multisensory weighting, robustness to changing conditions, and uncertainty calibration. Such functional comparisons can be informative without assuming equivalent biological mechanisms or subjective experience. Human studies of reliability-weighted cue combination provide benchmarks for evaluating whether a machine's sensor-fusion behavior follows specified computational expectations Ernst & Banks (2002) Alais & Burr (2004).
Artificial systems can be tested with version-controlled inputs and repeated perturbations, allowing systematic analysis of stimulus, model, and context effects. A sensor-fusion system may exhibit a tradeoff between precision, latency, and robustness that differs from the profile observed in a human task. A perceptual model trained on a narrow distribution can also fail under shifts in lighting, acoustics, or object categories. Comparing such systems requires matching task definitions and error metrics, reporting model versions, and testing on unseen environments. A superficial resemblance in one illusion or judgment does not establish identical representations or computational pathways Zhu et al. (2026).
The Artificial Era extends the research program by making variation across human and engineered perceivers a substantive comparative question. Human perception is situated in bodily development, active exploration, affect, and social participation; engineered sensing depends on technical architectures and training or calibration regimes. Joint experiments can study when differing perceptual profiles support cooperation, accessibility, or mutual error correction. They can also reveal cases where average-system design excludes minority human profiles. A principled comparative science specifies observable abilities and interaction outcomes, while philosophical interpretations of Artificial remain conceptually distinct from experimentally established psychological results Mollon et al. (2017) Zhu et al. (2026).
Conclusion
Individual differences in sensory perception are organized patterns of variation across receptors, neural processes, perceptual inference, learning, decision criteria, and environmental experience. Evidence from color appearance, auditory judgments, tactile estimates, bitter taste, odorant receptors, and multisensory experiments demonstrates that perceptual differences are real, often stimulus-specific, and sometimes biologically traceable. The defining scientific advance is the ability to model these differences as uncertain, multidimensional observer parameters rather than deviations from an average. Molecular findings can explain selected stimulus responses, while psychophysical and contextual analyses account for the wider functional organization of experience Emery & Webster (2019) Bufe et al. (2005) Keller et al. (2007).
Three methodological conclusions govern the field. First, a stable individual score requires reliable measurement across the relevant trials, stimuli, and sessions. Second, sensitivity must be separated from decision criteria, affective valuation, and self-reported environmental responsiveness. Third, group averages require analysis of variation within groups, especially when applied to development, autism, ADHD, and cultural comparisons. The contemporary evidence concerning audiovisual illusion stability, trait-measurement limits, and domain-sensitive autism research supports these priorities. Explanations become stronger when baseline modalities are tested independently and computational models are evaluated through predictive and parameter-recovery checks Hedge et al. (2018) Basu Mallick et al. (2015) Hadad et al. (2026).
Future research should establish representative multimodal normative datasets, longitudinal trajectories, cross-cultural measurement invariance, and robust links between laboratory performance and everyday participation. Mechanistically informative experiments can combine receptor biology, calibrated stimuli, adaptive testing, hierarchical models, and intervention studies. Comparative work with artificial perceptual systems adds a controlled domain for testing general principles of reliability, cue weighting, adaptation, and error correction. The resulting science treats sensory diversity as a core feature of perception and a foundation for designing environments and technologies that accommodate different observers. Its conclusions remain anchored in replicable evidence, explicit constructs, and context-sensitive interpretation Mollon et al. (2017) Zhu et al. (2026).
References
- Alais, D., & Burr, D. (2004). The ventriloquist effect results from near-optimal bimodal integration. Current Biology, 14(3), 257–262. https://doi.org/10.1016/j.cub.2004.01.029
- Aron, E. N., & Aron, A. (1997). Sensory-processing sensitivity and its relation to introversion and emotionality. Journal of Personality and Social Psychology, 73(2), 345–368. https://doi.org/10.1037/0022-3514.73.2.345
- Basu Mallick, D., Magnotti, J. F., & Beauchamp, M. S. (2015). Variability and stability in the McGurk effect: Contributions of participants, stimuli, time, and response type. Psychonomic Bulletin & Review, 22(5), 1299–1307. https://doi.org/10.3758/s13423-015-0817-4
- Bufe, B., Breslin, P. A. S., Kuhn, C., Reed, D. R., Tharp, C. D., Slack, J. P., Kim, U.-K., Drayna, D., & Meyerhof, W. (2005). The molecular basis of individual differences in phenylthiocarbamide and propylthiouracil bitterness perception. Current Biology, 15(4), 322–327. https://doi.org/10.1016/j.cub.2005.01.047
- Dunn, W., & Westman, K. (1997). The sensory profile: The performance of a national sample of children without disabilities. American Journal of Occupational Therapy, 51(1), 25–34. https://doi.org/10.5014/ajot.51.1.25
- Emery, K. J., & Webster, M. A. (2019). Individual differences and their implications for color perception. Current Opinion in Behavioral Sciences, 30, 28–33. https://doi.org/10.1016/j.cobeha.2019.05.002
- Ernst, M. O., & Banks, M. S. (2002). Humans integrate visual and haptic information in a statistically optimal fashion. Nature, 415(6870), 429–433. https://doi.org/10.1038/415429a
- Gori, M., Del Viva, M., Sandini, G., & Burr, D. C. (2008). Young children do not integrate visual and haptic form information. Current Biology, 18(9), 694–698. https://doi.org/10.1016/j.cub.2008.04.036
- Greven, C. U., Lionetti, F., Booth, C., Aron, E. N., Fox, E., Schendan, H. E., Pluess, M., Bruining, H., Acevedo, B., Bijttebier, P., & Homberg, J. (2019). Sensory processing sensitivity in the context of environmental sensitivity: A critical review and development of research agenda. Neuroscience & Biobehavioral Reviews, 98, 287–305. https://doi.org/10.1016/j.neubiorev.2019.01.009
- Hadad, B.-S., Leon, I., Hartston, M., Freud, E., & Abu-Akel, A. (2026). Mapping sensory sensitivity in autism. Molecular Autism, 17, Article 30. https://doi.org/10.1186/s13229-026-00719-y
- Hedge, C., Powell, G., & Sumner, P. (2018). The reliability paradox: Why robust cognitive tasks do not produce reliable individual differences. Behavior Research Methods, 50(3), 1166–1186. https://doi.org/10.3758/s13428-017-0935-1
- Jurek, L., Duchier, A., Gauld, C., Hénault, L., Giroudon, C., Fourneret, P., Cortese, S., & Nourredine, M. (2025). Sensory processing in individuals with attention-deficit/hyperactivity disorder compared with control populations: A systematic review and meta-analysis. Journal of the American Academy of Child & Adolescent Psychiatry, 64(10), 1132–1147. https://doi.org/10.1016/j.jaac.2025.02.019
- Keller, A., Zhuang, H., Chi, Q., Vosshall, L. B., & Matsunami, H. (2007). Genetic variation in a human odorant receptor alters odour perception. Nature, 449(7161), 468–472. https://doi.org/10.1038/nature06162
- Kim, U.-K., Jorgenson, E., Coon, H., Leppert, M., Risch, N., & Drayna, D. (2003). Positional cloning of the human quantitative trait locus underlying taste sensitivity to phenylthiocarbamide. Science, 299(5610), 1221–1225. https://doi.org/10.1126/science.1080190
- Leek, M. R. (2001). Adaptive procedures in psychophysical research. Perception & Psychophysics, 63(8), 1279–1292. https://doi.org/10.3758/BF03194543
- Mollon, J. D., Bosten, J. M., Peterzell, D. H., & Webster, M. A. (2017). Individual differences in visual science: What can be learned and what is good experimental practice? Vision Research, 141, 4–15. https://doi.org/10.1016/j.visres.2017.11.001
- Oxenham, A. J. (2018). How we hear: The perception and neural coding of sound. Annual Review of Psychology, 69, 27–50. https://doi.org/10.1146/annurev-psych-122216-011635
- Pluess, M., Aron, E. N., Kähkönen, J. E., Lionetti, F., Huang, Y., Tillman, T., Greven, C. U., & Aron, A. (2026). Evolution of the concept of sensory processing sensitivity and its measurement: The Highly Sensitive Person scale-revised. Personality and Individual Differences, 257, Article 113782. https://doi.org/10.1016/j.paid.2026.113782
- Robertson, C. E., & Baron-Cohen, S. (2017). Sensory perception in autism. Nature Reviews Neuroscience, 18(11), 671–684. https://doi.org/10.1038/nrn.2017.112
- Strand, J., Cooperman, A., Rowe, J., & Simenstad, A. (2014). Individual differences in susceptibility to the McGurk effect: Links with lipreading and detecting audiovisual incongruity. Journal of Speech, Language, and Hearing Research, 57(6), 2322–2331. https://doi.org/10.1044/2014_JSLHR-H-14-0059
- Wichmann, F. A., & Hill, N. J. (2001a). The psychometric function: I. Fitting, sampling, and goodness of fit. Perception & Psychophysics, 63(8), 1293–1313. https://doi.org/10.3758/BF03194544
- Wichmann, F. A., & Hill, N. J. (2001b). The psychometric function: II. Bootstrap-based confidence intervals and sampling. Perception & Psychophysics, 63(8), 1314–1329. https://doi.org/10.3758/BF03194545
- Zhu, H., Beierholm, U., & Shams, L. (2026). Analyzing multisensory integration: Do’s and don’ts. Neuroscience & Biobehavioral Reviews, 191, Article 107008. https://doi.org/10.1016/j.neubiorev.2026.107008