The Visual Scanning of Faces in Infants – Philip Salapatek

The investigation of early human ontogeny underwent a profound paradigm shift during the middle of the twentieth century, transitioning from speculative psychoanalytic interpretations and passive behaviorist frameworks to an empirical, psychophysically grounded developmental cognitive neuroscience. Central to this epistemological transformation was the pioneering experimental work of Philip Salapatek, whose micro-behavioral inquiries into how neonates and young infants visually navigate, segment, and encode their spatial environment unlocked unprecedented insights into the architecture of the preverbal mind. Prior to Salapatek’s rigorous ocular-motor investigations, the prevailing philosophical and clinical consensus largely conceptualized the human newborn as a functionally blind or sensory undifferentiated organism, overwhelmed by incoming stimulation and incapable of endogenous, rule-governed information extraction.

Salapatek fundamentally destabilized these historical assumptions by developing and refining high-precision corneal reflection photographic techniques capable of measuring the infant’s point of gaze with millimeter-level spatial fidelity. Through systematic, empirical demonstrations, he showed that far from wandering aimlessly or passively reacting to diffuse sensory shocks, the infant oculomotor system operates under defined, endogenously generated visual scanning rules. The apex of this scientific contribution crystallized in his landmark investigations of facial exploration—most notably synthesized in his monumental 1975 monograph and associated empirical studies—which chronicled the qualitative developmental shift that occurs between the first and second months of human life. In tracking the exact coordinates of infant fixations across two-dimensional geometric forms and human facial stimuli, Salapatek provided the empirical architecture for what would become known as the “externality effect” and illuminated the biological timetable governing the infant’s transition from reflexive edge-detection to holistic, socially motivated facial processing.

This comprehensive treatise examines Philip Salapatek’s foundational research on infant visual processing, mapping its biographical, theoretical, and technical antecedents, dissecting the experimental mechanics of corneal reflection oculometry, and detailing the psychophysical dynamics of early visual exploration. In tracing the progression from subcortical, retinotectal visual guidance to cortical, geniculostriate-driven feature integration, this analysis explores the structural, computational, and social dimensions of Salapatek’s work. His empirical legacy established the infant as an active, computational information gatherer, directly anticipating contemporary discoveries in developmental neuroimaging, evolutionary psychology, computational vision, and the early detection of neurodevelopmental conditions such as autism spectrum disorder.

1. Introduction to Philip Salapatek’s Foundational Research on Infant Visual Processing

The systematic study of infant visual perception requires balancing technological innovation with empirical psychophysics. Prior to the late 1960s, the visual world of the human infant was largely a matter of theoretical conjecture rather than direct empirical measurement. Philip Salapatek’s entry into this domain marked a critical turning point, replacing indirect behavioral inferences with precise spatial and temporal measurements of visual gaze, thereby establishing the foundations of modern infant ocular psychophysics.

1.1 Biographical Context and the Experimental Milieu of the 1960s and 1970s

Philip Salapatek completed his doctoral training at Yale University under the mentorship of William Kessen, an influential developmental psychologist whose laboratory helped transform infant research from descriptive child study into a rigorous experimental discipline. During the mid-1960s, experimental psychology was experiencing the initial tremors of the “cognitive revolution.” The long-standing dominance of radical behaviorism—which viewed the preverbal infant as a passive stimulus-response mechanism that could only be conditioned through overt motor acts—was yielding to models emphasizing active mental representations, sensory filtering, and internal cognitive structures. However, while adult cognitive psychology rapidly adopted information-processing paradigms based on verbal reports, reaction times, and memory recall, the developmental psychologist studying preverbal human neonates faced a severe methodological impasse: how could one rigorously evaluate the cognitive architecture of an organism that possessed neither speech nor coordinated voluntary motor reaching?

Salapatek recognized that the ocular motor system provided a viable behavioral window into the internal processing of the preverbal infant. At Yale, Kessen and Salapatek sought to establish laboratory controls comparable to adult psychophysics while accounting for the biological vulnerability and rapid state fluctuations of the infant. The research atmosphere was charged with the ambition to quantify the human sensory apparatus at its absolute ontogenetic onset. Rather than relying on broad observational protocols or parent logs, Salapatek brought infants directly into controlled optical laboratories, engineering specialized apparatuses to measure visual tracking with photographic precision. This experimental milieu provided the intellectual springboard for Salapatek’s career, leading to his foundational appointments at the University of Minnesota’s Institute of Child Development, where he continued to refine the mechanical and theoretical tools required to map the micro-ecology of the human infant’s visual gaze.

1.2 The Seminal Inquiry: How Preverbal Infants Extract Information from Human Faces

Within the sensory landscape of the infant, the human face stands out as an ecological stimulus rich in structural complexity, dynamic movement, affective charge, and evolutionary relevance. For Salapatek, the human face served as the ultimate test case for investigating the rules of preverbal information processing. The primary question driving his work was deceptively straightforward yet methodologically profound: when a human newborn looks at an adult face, what does that infant actually see, and how is visual information gathered to form a coherent percept? Salapatek sought to determine whether an infant’s visual engagement with a face was governed by a generalized, passive ocular capture driven strictly by peripheral luminance changes, or whether it reflected an active, endogenous visual search program designed to sample specific internal features carrying psychological and communicative value.

To resolve this question, it was necessary to differentiate between global pattern awareness and fine-grained, localized foveal fixation. It had already been demonstrated that infants could broadly discriminate patterned stimuli from uniform fields, but these observations left the precise micro-structure of infant gaze unknown. Salapatek hypothesized that visual scanning underwent an organized developmental transition over the first twelve weeks of life. He posited that the visual apparatus of the one-month-old infant was structurally constrained, limiting their ability to process complex internal configurations, whereas the two- to three-month-old infant experienced a qualitative cognitive and neural reorganization. This transformation, he theorized, liberated the gaze from outer physical boundaries and directed it toward internal features—such as the eyes and mouth—that facilitate social attachment, affective communication, and identity recognition.

1.3 Overcoming Methodological Barriers in Studying Preverbal Populations

Conducting optical psychophysics with preverbal infants involves unique methodological and physiological challenges. Visual acuity in the newborn is poor—often estimated between 20/400 and 20/800 Snellen equivalents—and the infant’s contrast sensitivity function is significantly attenuated, shifted toward extremely low spatial frequencies. Consequently, an optical stimulus that appears rich in fine detail to an adult observer may present as an indistinct, washed-out blur to a four-week-old infant. Furthermore, nonverbal subjects cannot follow instructional sets to calibrate their gaze to optical coordinates, nor can they be subjected to prolonged experimental trials without rapid fatigue, fussiness, crying, or falling into deep sleep states.

Beyond these physiological limitations, Salapatek faced the challenge of distinguishing passive, reflexogenic ocular-motor capture—mediated by subcortical structures like the superior colliculus—from active, endogenous exploratory gaze driven by developing cortical circuits. Historically, developmental researchers had measured macroscopic head-turning or relied on broad observer judgments of an infant’s general head orientation toward a visual target. These gross measures could not determine whether an infant fixating a face was inspecting the pupil of the eye, tracing the border of the temporal hairline, or simply caught in an obligatory gaze lock on a point of high peripheral contrast. Salapatek understood that resolving this question required moving past coarse behavioral metrics to millimeter-scale foveal positioning, capturing visual fixations at temporal resolutions high enough to plot individual saccadic vectors without causing distress to the infant.

2. Historical and Theoretical Context of Infant Vision in the Mid-20th Century

To understand the paradigm shift initiated by Salapatek’s work, one must examine the theoretical landscape of perceptual development in the early and mid-20th century. This era was characterized by a sharp divide between empiricist models of sensory passivity and emerging nativist formulations, set against early psychophysical methodologies that struggled to verify internal mental states.

2.1 The Tabula Rasa and Sensory Blooming Confusion Traditions

For decades, Western psychology’s conceptualization of early infant perception was dominated by the empiricist assertion popularized by William James in The Principles of Psychology (1890), which posited that the neonate is assaulted by an undifferentiated influx of sensory data, experiencing the external world as one “blooming, buzzing confusion.” This perspective aligned with the philosophical doctrine of the tabula rasa, which asserted that the mind arrives in the world devoid of internal organization, structural schema, or visual processing rules. According to this view, the sensory organs of the newborn merely registered raw physical energies, leaving the higher perceptual organization of objects, depth, and social stimuli to be gradually constructed through months of tactile, motor, and associative learning.

This empiricist view was further reinforced by early psychoanalytic frameworks and behaviorist theories, both of which operated on the assumption of neonatal sensory passivity and functional blindness. Behaviorists argued that because internal representations could not be directly observed or conditioned via classical paradigms in the first weeks of life, the neonate was largely a reflexive, decorticate-like organism. Early medical psychophysics similarly maintained that because the human cortex—specifically the primary visual cortex (striate area V1)—is histologically immature at birth, the newborn must be functionally subcortical, governed entirely by lower-order visual reflexes lacking cognitive organization. This conceptual framework created a widespread belief that studying fine-grained visual strategies in the first two months of life was largely futile, as the infant was presumed incapable of executing structured perceptual processing.

2.2 Robert Fantz and the Preferential Looking Paradigm

This consensus began to unravel in the late 1950s and early 1960s with the groundbreaking work of Robert L. Fantz. Fantz challenged the concept of the passive neonate by introducing the preferential looking technique, a behavioral method designed to assess infant visual discrimination without requiring verbal instruction or motor training. Fantz placed infants in a specialized viewing chamber (the “looking chamber”) where they were simultaneously presented with two visual patterns side by side—such as a series of concentric circles versus plain stripes, or a schematic face pattern versus a scrambled arrangement of the same features. By observing the reflection of the visual targets on the infant’s cornea through a hidden peephole, an experimenter could calculate the aggregate dwell time dedicated to each stimulus.

Fantz’s experiments demonstrated that even neonates looked significantly longer at patterned, complex, and face-like figures than at uniform, unpatterned surfaces, proving that the human visual system possesses innate or rapidly emerging pattern preferences. However, while Fantz’s paradigm delivered a critical empirical blow to the tabula rasa model, it possessed an inherent methodological ceiling: macro-level preference (total looking time) could not reveal the underlying micro-level perceptual mechanisms. For example, knowing that an infant looks at a schematic face for twenty out of thirty seconds does not explain how the visual information was processed. Did the infant take in the entire facial pattern holistically? Did they fixate exclusively on an anomalous edge, or did their gaze systematically bounce between the eyes and the mouth? Salapatek recognized that total fixation duration was an aggregate metric that obscured the real-time temporal and spatial dynamics of visual information processing.

2.3 Nativist versus Empiricist Frameworks of Early Cognitive Development

The gap between Fantz’s macro-level preferential looking and the micro-structure of perception intensified theoretical debates regarding early cognitive development. On one side, nativist theorists argued that the human visual system arrives pre-configured with specialized representations. In the social domain, this implied an innate mental representation or module specifically tuned to detect conspecifics, allowing the infant to orient toward faces as ecologically essential biological entities. This perspective was closely tied to James J. Gibson’s ecological optics, which proposed that perceptual systems evolved to extract invariants and directly perceive environmental “affordances”—such as the social affordance of an adult’s communicative gaze—without needing to slowly build mental representations through sensory association.

In contrast, constructivist and empiricist traditions, tracing back to Jean Piaget, argued that perceptual organization is built through the continuous coordination of sensorimotor schemes. From this perspective, an integrated percept of an object or face cannot exist a priori; it must be constructed via circular reactions and sensorimotor interactions over time. Salapatek’s micro-behavioral scanning investigations occupied a rigorous empirical middle ground within this debate. Rather than committing entirely to pure nativism or pure empiricism, Salapatek adopted an approach rooted in structural psychophysics: he sought to determine how biological hardware—specifically, the anatomical maturation of the retina, the subcortical visual centers, and the emerging striate cortex—mechanically constrained and channeled the infant’s information gathering. By systematically analyzing the spatial coordinates of visual fixations, Salapatek revealed that early perception was neither a formless blooming confusion nor an adult-like mental representation, but rather a structured sequence of biological constraints undergoing rapid neuromotor maturation.

3. Methodological Framework: Infrared Corneal Reflection and Eye-Tracking Apparatus

To move past the limits of preferential looking, Salapatek had to pioneer an experimental platform capable of resolving the exact spatial coordinates of an infant’s foveal gaze. This led to the development of the infant corneal reflection eye-tracking apparatus, an engineering achievement that established the methodological standard for developmental psychophysics in the late 20th century.

3.1 Principles of the Corneal Reflection-Pupil Center Technique

The core optical physics behind Salapatek’s experimental methodology relied on the geometry of the corneal reflection-pupil center technique. When an external light source illuminates the eye, it reflects off the anterior convex surface of the cornea, acting as a convex spherical mirror. This primary reflection is known as the first Purkinje image. Because the radius of curvature of the cornea is smaller than the radius of rotation of the eyeball itself, when the eye rotates to fixate on different points in the visual field, the relative spatial distance between the center of the pupil and the corneal reflection shifts systematically and predictably.

To implement this technique with vulnerable human infants, Salapatek used harmless infrared illumination. Using wavelengths in the near-infrared spectrum offered two distinct methodological advantages:

• Infrared light was invisible to the infant, avoiding retinal dazzle, discomfort, or visual distraction from the experimental targets.
• It prevented the infant’s pupils from constricting, maintaining a dilated pupillary aperture that made the pupillary margin stand out in high contrast against the darker iris and surrounding sclera.

By recording this optical interaction, Salapatek could apply mathematical algorithms that calibrated the angular displacement of the corneal glint relative to the pupillary center, calculating the infant’s instantaneous visual line of regard across an external stimulus plane with high spatial accuracy.

3.2 Apparatus Architecture and Experimental Stabilization

Constructing an apparatus capable of tracking infant gaze required isolating the eye’s rotation from extraneous head and bodily movements. An adult participant in an eye-tracking experiment can be instructed to bite down on a dental impression plate or rest within a rigid chin-rest assembly. However, applying these mechanical restraints to an infant produces distress, crying, and immediate behavioral shutdown, invalidating the psychophysical data.

Salapatek resolved this dilemma by designing a specialized viewing chamber, often referred to as the Minnesota Infant Eye-Movement Apparatus. The infant was placed in an inclined, ergonomically contoured cradle that supported the spine and neck. To reduce head movement without causing distress, the cradle used soft lateral pads that gently bounded the infant’s temporal regions, dampening gross rotational and translational movements while allowing natural respiration and minor postural shifts. Above the cradle, a light-tight enclosure housed a system of front-surface optical split-mirrors, telephoto camera lenses, and infrared light sources. The split-mirror configuration allowed an infrared-sensitive camera to capture an unobstructed, high-magnification image of the infant’s eye, while the infant looked directly through the mirror at a visual stimulus positioned above them. Ambient illumination, target luminance, and stimulus distance were standardized to ensure that visual contrast remained constant throughout the testing session.

3.3 Data Recording, Frame-by-Frame Cinematography, and Spatial Mapping

Decades before the advent of digital charge-coupled devices (CCD), high-speed computing, and automated computer-vision algorithms, recording and analyzing infant gaze required labor-intensive analog data processing. The visual output of Salapatek’s apparatus was recorded on high-contrast 16mm or 35mm cinematographic film at standard rates of 8, 16, or 24 frames per second. Each physical frame had to be individually developed, projected onto a grid, and analyzed by trained laboratory coders.

The manual data analysis process was exacting:

• Coders measured the exact horizontal and vertical coordinates of the corneal glint relative to the pupillary boundaries frame-by-frame.
• These measurements were translated through calibration matrices into two-dimensional spatial coordinates on the target stimulus plane.
• Sequential fixations were combined into saccadic vectors, fixational clusters, and spatial dwell times across the stimulus.
• Rigorous inter-rater reliability protocols were maintained, requiring multiple independent coders to classify each frame blindly.
• Mathematical dispersion indices were calculated to quantify whether an infant’s gaze was concentrated tightly within a small visual region or broadly dispersed across the stimulus display.

4. The Externality Effect: Edge Detection and Saccadic Limitations in Neonates

The primary empirical breakthrough resulting from this methodology was the discovery and characterization of the externality effect. This phenomenon, which Salapatek identified and analyzed in the late 1960s and 1970s, provided empirical evidence that the newborn visual system processes complex two-dimensional visual scenes through structured, rule-governed constraints.

4.1 Defining the Externality Effect in One-Month-Old Infants

The externality effect describes an ocular-motor phenomenon observed in human neonates and one-month-old infants: when presented with a complex, compound visual figure consisting of an outer contour containing one or more internal elements, the young infant’s visual scanning is heavily restricted to the outermost boundary of the figure. The infant frequently fails to bring their fovea into the interior of the stimulus, effectively ignoring internal features that are easily visible to older infants.

Salapatek demonstrated this phenomenon using both two-dimensional geometric figures—such as an internal square or circle nested within a larger external triangle—and naturalistic human faces. When one-month-old infants were shown a composite geometric shape, their fixations clustered almost entirely along the high-contrast perimeter edges. The infants scanned back and forth across outer vertices and linear segments, rarely executing the centripetal saccades required to cross the internal void and sample the nested figure. The high-contrast perimeter appeared to serve as a visual barrier, capturing the infant’s ocular motor system and preventing the spatial exploration of internal details.

4.2 Oculomotor Immature Mechanisms and Subcortical Guidance

Salapatek did not view the externality effect merely as an isolated behavioral quirk, but rather as the behavioral readout of an immature neuroanatomical visual system. In the first weeks of life, visual orienting is predominantly controlled by subcortical neural structures, most notably the superior colliculus and the retinotectal pathway. The evolutionary role of the superior colliculus is to detect gross, transient physical changes in the visual periphery and trigger rapid, reflexive saccades that align the fovea with these novel, high-contrast energetic shifts.

However, the young infant’s visual system lacks the mature cortical architecture necessary to regulate this reflexive drive. The cortical networks—specifically the primary visual cortex (V1), extrastriate areas, and the frontal eye fields (FEF)—that execute top-down inhibitory control over subcortical reflexes are structurally immature at birth. As a result, the one-month-old infant experiences what developmental neuroscientists term “sticky fixation.” Once the collicular pathway detects and locks onto a high-contrast external boundary (such as the sharp luminance transition between an outer border and a dark background), the infant lacks the cortical inhibition needed to disengage from that edge. This immature state is compounded by the neonate’s low contrast sensitivity thresholds, which are tuned primarily to low spatial frequencies. Because an external perimeter typically presents the largest, highest-contrast spatial boundary in the visual field, it dominates the subcortical motor response, trapping the infant’s gaze along the outer margins.

4.3 Exceptions and Boundary Conditions of the Externality Effect

Through systematic experimental manipulations, Salapatek and his contemporaries identified important boundary conditions and exceptions to the externality effect. The edge-bound scanning pattern was not an absolute mechanical limitation; rather, it was a probabilistic bias that could be overridden by specific physical inputs. Salapatek observed that if an internal element within a compound figure moved while the outer perimeter remained static, the one-month-old infant could successfully disengage from the edge and redirect their fovea toward the internal target. The temporal modulation introduced by physical movement effectively provided a strong sensory signal that subcortical circuits could prioritize over the static outer border.

Similarly, internal features could break the externality trap if their physical salience was substantially heightened. When an internal element flashed, modulated its luminance rapidly, or featured a much higher local contrast ratio than the surrounding boundary, the infant’s gaze could penetrate the interior of the figure. Structural geometry also imposed boundary constraints: if an internal element was positioned unusually close to the outer perimeter—falling within the spatial window of a single, small saccade—or if the overall size of the stimulus was reduced to fit within the infant’s immediate foveal window, edge trapping was reduced. These nuanced observations showed that the externality effect was not simply a blind sensory deficit. Instead, it was an interaction between low-level visual salience, stimulus geometry, and developing cortical control over oculomotor disengagement.

5. The Developmental Shift: Comparing One-Month-Old and Two-Month-Old Scanning Behaviors

The core of Salapatek’s 1975 monograph lies in his empirical documentation of the developmental transition occurring between four and eight weeks postnatal. This brief window marks a significant milestone in infant visual development, representing a qualitative transition from subcortically dominated reflexes to flexible, cortically mediated visual exploration.

5.1 The Qualitative Transition Between Four and Eight Weeks Postnatal

Between four and eight weeks after birth, the human infant undergoes a major functional reorganization of their visual scanning behavior. Salapatek’s quantitative kinematic data demonstrated that this period is characterized by more than just minor, incremental improvements in acuity; it represents an overhaul of visual exploratory mechanics. In one-month-old infants, visual fixations are typically long, prolonged, and tightly clustered along outer edges, punctuated by clumsy, inaccurate saccades that often miss their targets and require corrective re-orientations.

By eight to ten weeks of age, this rigid scanning pattern changes markedly. Average fixation durations decrease as saccadic frequencies increase, indicating more efficient visual information processing. Saccadic amplitudes become more flexible, with infants showing greater fixational stability once a target is acquired. Most importantly, the rigid gaze lock of the externality effect diminishes. Two-month-old infants display the ability to disengage their gaze from high-contrast perimeters, systematically sweeping their fovea into the interior of both geometric and biological targets. This marks the transition from isolated, univariate fixations to integrated, multivariate visual exploration, allowing the infant to sample multiple components of a visual scene within a single viewing period.

5.2 Internal Feature Penetration: The Eye and Mouth Fixation Milestone

This functional reorganization is clearly visible in the infant’s scanning of the human face. When Salapatek recorded the gaze of one-month-old infants presented with real or photographic faces, the scanpaths were dominated by peripheral tracking: their fixations tracked the outer contour of the chin, lingered along the temporal hairline, or clustered around the external curves of the ears. The internal regions of the face—where identity, intention, and emotional expressions are communicated—were mostly left unvisited.

By two months of age, this scanning pattern shifts inward. In his 1975 studies, Salapatek observed that the two-month-old infant’s gaze moves systematically into the interior of the face, concentrating primarily on the eyes and mouth:

• The hairline and outer boundaries become secondary reference points, with total fixation time on the internal facial core increasing dramatically.
• When the stimulus face engaged in vocalization or mouth movements, the two-month-old infant rapidly shifted visual attention between the eyes and the moving lips.
• This internal feature penetration coincides directly with the emergence of the social smile—the developmental milestone around six to eight weeks when an infant smiles back at an adult caregiver.

Salapatek’s data showed that the social smile is not a generic response triggered by a vague visual percept, but is closely tied to the oculomotor ability to look directly at the internal, communicative features of the human face.

5.3 Individual Differences and Developmental Trajectories

While this developmental transition generally peaks between six and eight weeks postnatal, Salapatek documented individual variations across cohorts. He noted that chronological age from birth was often a less reliable predictor of internal feature penetration than post-conceptional age (gestational age at birth plus weeks of postnatal life). Preterm infants evaluated at eight weeks postnatal, for example, often retained the edge-bound scanning characteristics of younger term infants, indicating that this visual transition depends directly on biological neurological maturation rather than postnatal visual experience alone.

Furthermore, Salapatek’s work connected these scanning shifts to broader physiological indicators of cognitive attention, such as autonomic heart-rate deceleration. He observed that internal feature exploration was systematically accompanied by sustained cardiovascular deceleration, an established marker of active, sustained cognitive processing. Conversely, the long, sticky fixations of the one-month-old caught on an external perimeter were often accompanied by autonomic recovery or signs of distress, indicating that the young infant was visually trapped rather than cognitively engaged. Tracking individual scanning patterns over time confirmed that the shift from external to internal scanning is a stable, non-random developmental trajectory shared across typically developing infants.

6. Micro-Analysis of Facial Exploration: Fixation Distribution Across Anatomical Landmarks

Salapatek’s detailed spatial mapping provided the first empirical micro-analysis of how infants explore the primary anatomical landmarks of the human face. By moving past broad measures of total dwell time, his work established which specific facial structures anchor visual attention during early infancy.

6.1 The Primacy of the Eyes: Morphological High-Contrast and Social Signalling

Within Salapatek’s micro-analyses of two-month-old scanning, the eye region emerged as the primary visual anchor of the human face. When two-month-olds were presented with static facial stimuli, the upper facial quadrant—specifically the ocular region—received the highest density of visual fixations. Fixations were not simply cast broadly toward the upper half of the head; they targeted the eyes themselves with high spatial precision.

Salapatek explained this bias through an interaction of structural and communicative factors:

• Structural Salience: Morphologically, the human eye is uniquely high in visual contrast. The dark, rounded boundary of the iris set against the bright white background of the surrounding sclera, framed by the upper eyelid and eyebrow, forms an ideal target for an infant visual system tuned to low and intermediate spatial frequencies.
• Communicative Alignment: This structural contrast naturally directs the infant’s gaze to the primary region for social signaling. The eyes convey vital communicative information, including shifts in focus, social intent, and affective states.

Salapatek observed that the infant’s ocular motor system was particularly responsive to direct versus averted gaze. When an adult face established direct eye contact with the infant, fixation clusters on the eyes became tighter and were maintained for longer durations. When the eyes looked away, the infant’s gaze became more dispersed, often sliding down toward secondary features or drifting off the stimulus entirely. This confirmed that the human infant’s visual scanning is attuned to detect and maintain mutual social contact.

6.2 Scanning of the Mouth, Nose, and Facial Contours

While the eyes served as the primary anchor for visual attention, Salapatek mapped distinct scanning patterns across other anatomical landmarks, noting variations tied directly to movement and contrast:

• The Mouth: In static photographic presentations, the mouth received significantly fewer fixations than the eyes, but remained a prominent secondary target. However, this dynamic changed when the stimulus became active: during live speech or when the lips formed dynamic shapes, the two-month-old infant redirected their gaze to the mouth region, often tracking the motion of the lips before returning to the eyes.
• The Nose: The nose was largely neglected by the infant’s visual scan. Salapatek’s spatial maps showed an absence of fixation clusters over the nasal bridge and tip. Because the nose lacks sharp luminance transitions or high-contrast borders—relying instead on soft shadows and gradual depth contours—it failed to trigger the infant’s contrast-dependent oculomotor system.
• Facial Contours: The chin, temporal hairline, and lateral perimeter of the head remained occasional reference points for the two-month-old, but were no longer the traps seen in one-month-olds. Instead, the infant used these peripheral landmarks as brief transit points between exploratory passes across the internal features.

6.3 The Triangular Scanpath Model and Relational Geometry

By plotting consecutive saccades, Salapatek observed that visual fixations were not scattered randomly across the face; rather, they followed an organized spatial trajectory. This led to the formulation of what is now recognized as the triangular scanpath model. When a two-month-old infant explores an upright human face, their saccadic sweeps frequently trace an inverted isosceles triangle, connecting the left eye, the right eye, and the mouth.

This triangular scanpath revealed that the two-month-old infant was beginning to process the structural geometry of the face. The visual system was not just identifying individual features in isolation; it was actively mapping the relational distances between them. Salapatek demonstrated this by presenting infants with inverted (upside-down) faces. When inverted, this organized triangular scanning was disrupted: saccadic paths became irregular, fixations scattered, and the infant struggled to establish a coherent scanning route. This showed that the infant visual system is attuned to the canonical, upright configuration of the human face, laying the groundwork for adult-like configural and holistic facial processing.

7. Structural vs. Social Determinants of Infant Visual Attention

The discovery that infants systematically scan the eyes and mouth led to an important theoretical debate in developmental cognitive psychology: does the infant scan faces because they possess an innate social understanding of conspecifics, or is this behavior the mechanical consequence of low-level physical features like contrast, luminance, and spatial frequency?

7.1 The Physicalist Hypothesis: Salience, Contrast, and Spatial Frequency

The physicalist perspective argued that early visual scanning could be explained without invoking specialized social or cognitive modules. Proponents of this view contended that infants look at the eyes and outer contours simply because these regions contain the highest physical contrast and edge density in the human face. The infant visual system functions as a low-level physical filter, characterized by its Modulation Transfer Function (MTF) and contrast sensitivity profile.

Salapatek took this physicalist hypothesis seriously, systematically testing artificial stimuli matched for spatial frequency, edge density, and luminance contrast against real faces:

• Infants were presented with black-and-white checkerboards, geometric polygons, and high-contrast gratings to isolate the low-level properties that drive saccadic movements.
• The results confirmed that an infant’s gaze can be guided by low-level physical salience: whenever a non-social stimulus presented sharp, high-contrast borders, the one-month-old infant exhibited the same edge-trapping behavior observed when viewing human hairlines.
• Physical properties clearly constrained the infant’s gaze, establishing the spatial boundaries within which visual processing could occur.

7.2 The Social Pre-Wiring Hypothesis: Innate Face Templates

While low-level physical properties clearly shape infant vision, they cannot fully explain the infant’s selective responsiveness to the human face. This led to the development of the social pre-wiring hypothesis, synthesized by John Morton and Mark Johnson in their influential CONSPEC and CONLERN model, which drew heavily on Salapatek’s empirical findings. Morton and Johnson proposed a dual-system framework of early facial processing:

• CONSPEC: An early, subcortically mediated structural mechanism (operating primarily through the superior colliculus and pulvinar) that guides the infant’s visual attention toward stimuli matching a crude, innate template of conspecific faces—specifically, three high-contrast blobs arranged in a triangle (two eyes, one mouth) on an upright oval background.
• CONLERN: A later-developing, cortical learning mechanism that takes over around two to three months of age, driven by the geniculostriate pathway and ventral visual cortex. CONLERN processes the fine-grained, identity-specific details of individual faces, building long-term social representations.

Salapatek’s work provided the empirical link between these two systems. His observations demonstrated that the one-month-old is governed by low-level physical constraints (consistent with CONSPEC’s subcortical origins), while the two-month-old’s transition to internal facial features marks the initial emergence of CONLERN’s cortical processing, showing how general biological tuning mechanisms provide the foundation for rapid social learning.

7.3 Dynamic Versus Static Stimuli: The Impact of Movement and Voice

Salapatek recognized that in the real world, faces are rarely static, silent, two-dimensional images. Under natural conditions, the human face is an active, multi-modal source of continuous motion and vocal sound. To assess the ecological validity of his photographic findings, Salapatek integrated dynamic and multi-modal elements into his experimental setups.

These studies showed that internal motion—such as moving lips, shifting eyes, or raising eyebrows—fundamentally alters infant gaze dynamics. Movement acted as an overriding factor for the externality effect: even one-month-old infants, who remained edge-bound when viewing a static portrait, could break free from the perimeter to fixate a moving mouth. When human vocalization was added, producing an integrated audio-visual display, this internal fixation was reinforced. Infant gaze aligned closely with speech articulation, demonstrating that cross-modal sensory integration helps orient the infant visual system toward internal communicative features well before cortical networks reach full functional maturity.

8. Comparative Analysis: Geometric Figures vs. Live and Static Human Faces

To determine whether facial scanning was unique or part of a general visual processing program, Salapatek conducted comparative investigations contrasting how infants explore abstract geometric shapes versus live and static human faces. This comparative approach highlighted how the structural complexity of a visual target directly shapes infant oculomotor behavior.

8.1 Scanning Geometric Solids and Polygons (Salapatek & Kessen 1966, 1973)

Before publishing his 1975 facial monograph, Salapatek, in collaboration with William Kessen, conducted a series of studies examining how neonates explore simple geometric figures, such as solid triangles, nested circles, and varied polygons. In their 1966 study published in the Journal of Experimental Child Psychology, they demonstrated that human newborns do not sweep their gaze across an entire geometric shape to calculate its surface area or map its holistic form. Instead, the neonate’s gaze typically clustered on a single vertex or sharp corner of a triangle.

The vertex, representing a sharp direction change with high local contrast, was sufficient to capture and anchor the newborn’s oculomotor system. The infant would fixate on that single vertex for prolonged periods, making short, micro-saccades around that point while ignoring the remaining corners and connecting edges. When presented with nested figures (such as a small circle inside a large triangle), the infant scanned the outer boundary, rarely looking into the interior. These findings confirmed that single-feature capture and the externality effect were general properties of the immature visual system, rather than behaviors specific to human faces.

8.2 Scrambled, Inverted, and Schematic Face Paradigms

To isolate which properties of the face guide the transition to internal scanning, Salapatek and his contemporaries used schematic line drawings featuring systematically manipulated facial components. These included:

• Symmetrically Arranged Faces: Canonical schematic faces with internal features positioned in their natural locations.
• Scrambled Faces: Figures where the internal features (eyes, nose, mouth) were rearranged—such as placing the mouth in the forehead region and the eyes vertically along the cheek.
• Inverted Faces: Standard faces presented upside-down, disrupting normal spatial relationships while preserving identical local features.

The empirical results revealed a clear developmental dissociation:

• One-month-olds scanned both canonical and scrambled schematic drawings similarly, remaining caught on the outer margins regardless of internal feature arrangement.
• Two-month-olds, by contrast, responded selectively to structural changes. When viewing scrambled faces, their smooth triangular scanpaths broke down into fragmented, erratic fixations with increased dwell times, as if registering a structural anomaly.
• When presented with inverted faces, two-month-olds showed a marked reduction in dwell time on the eye region, accompanied by scanning patterns typical of abstract geometric shapes.

This confirmed that by eight to ten weeks of age, visual processing had moved beyond isolated feature detection to configural, holistic processing that expects facial components to align with canonical human geometry.

8.3 Representational Realism: From Line Drawings to Real Human Caregivers

Salapatek also evaluated how representational realism affects infant visual engagement, examining infant gaze across a spectrum of complexity: ranging from high-contrast line drawings and black-and-white photographs to natural color images and live, interactive human faces. His observations revealed that representational complexity shapes both attentional engagement and the stability of infant scanpaths.

While high-contrast line drawings elicited concentrated fixations on specific edges, naturalistic faces—complete with realistic skin tones, subtle depth cues, soft gradations, and natural contours—elicited broader, more varied visual exploration. When presented with their real mother versus a female stranger, two-month-old infants exhibited distinct scanning differences: fixations on the mother were characterized by longer periods of sustained gaze, reduced latency in locating the eyes, and lower scanning dispersion compared to stranger faces. This gradient of realism confirmed that while artificial stimuli are valuable for experimental control, naturalistic faces engage additional visual and emotional processing mechanisms that integrate depth, familiar luminance distributions, and socio-emotional recognition.

9. Neural Maturation and the Shift from Subcortical to Cortical Visual Pathways

The behavioral shift in infant visual scanning documented by Salapatek corresponds directly with the structural maturation of the underlying central nervous system. Modern developmental neuroscience views the transition from edge-bound to internal facial scanning as the behavioral marker of a major neuroanatomical shift: the passing of oculomotor control from primitive subcortical circuits to emerging cortical pathways.

9.1 Subcortical Architecture: The Retinotectal Pathway and Superior Colliculus

At birth, human visual function is predominantly mediated by the subcortical retinotectal pathway, an evolutionarily primitive visual processing circuit centered on the superior colliculus. The superior colliculus receives direct inputs from the retina (primarily from the peripheral fields) and projects directly to the brainstem motor generators that control saccadic eye movements. This subcortical system is specialized for rapid orienting toward peripheral transients, high-contrast borders, and sudden movement, functioning as an automated spatial alarm system.

However, the retinotectal pathway possesses limited spatial resolution, lacks chromatic sensitivity, and is incapable of decomposing complex visual patterns. Because subcortical circuits operate largely on gross luminance transitions, they are easily captured by the outermost boundary of an object, explaining the neuroanatomical origin of the externality effect. Furthermore, the early collicular reflex pathway lacks the top-down cortical inhibition required to voluntarily break fixation, resulting in the prolonged “sticky fixations” observed in one-month-olds. Until cortical connections mature and integrate with the collicular machinery, the young infant remains structurally biased toward peripheral edge-trapping.

9.2 Cortical Maturation: Geniculostriate Pathway and Primary Visual Cortex (V1)

The developmental transition observed at two months reflects the functional maturation of the geniculostriate pathway, which runs from the retina to the lateral geniculate nucleus (LGN) of the thalamus and on to the primary visual cortex (striate area V1). At birth, human V1 is structurally and functionally immature: dendritic arborization is sparse, synaptic density is low, and the myelination of the optic radiations is incomplete. Over the first eight weeks of postnatal life, an intense wave of synaptogenesis, laminar differentiation, and axonal myelination occurs within the striate cortex.

This anatomical maturation transforms infant visual capabilities:

• Cortical orientation columns and spatial-frequency filters become functionally operational, allowing the infant to resolve intermediate and high spatial frequencies.
• This enables the infant to process the fine internal details of visual scenes, such as the pupils, irises, and mouth contours.
• Simultaneously, descending projections from V1 and frontal regions begin to modulate subcortical activity, providing the inhibitory control needed to suppress collicular reflexes and voluntarily disengage from peripheral edges.

The behavioral consequence of this cortical maturation is the end of the externality effect and the emergence of flexible internal visual exploration.

9.3 The Emergence of the Ventral Stream and the Fusiform Face Area Precursors

As striate cortex matures, functional specialization begins to take shape along the two primary cortical visual pathways: the dorsal “where/how” stream and the ventral “what” stream. The ventral visual pathway—extending from V1 through visual area V4 to the posterior inferior temporal cortex—is responsible for object identification, structural pattern analysis, and facial recognition. In adults, facial processing relies on a dedicated functional network, most notably the Fusiform Face Area (FFA) and the Superior Temporal Sulcus (STS).

Neurodevelopmental models suggest that early precursors to the FFA and STS become functionally integrated during the second and third months of life. Salapatek’s behavioral findings show that the emergence of internal feature scanning provides the visual input required to calibrate these cortical areas. By repeatedly sweeping their fovea over the eyes, nose, and mouth, two-month-old infants expose their developing ventral stream to high-resolution visual input, establishing the retinotopic and category-specific representations needed for mature face-processing networks.

10. Methodological Replications, Divergences, and Contemporary Paradigms

Salapatek’s empirical findings provided the foundation for a wide body of developmental research throughout the late 20th and early 21st centuries. His core paradigms have been replicated, refined, and expanded by successive generations of developmental cognitive neuroscientists using increasingly sophisticated tracking systems.

10.1 Replication Efforts: Haith, Maurer, and Bushnell

Following Salapatek’s initial publications, several researchers set out to replicate and refine his findings under varying experimental conditions:

• Marshall Haith: Haith and colleagues investigated infant scanning in complete darkness using infrared systems, discovering that even without visual input, neonates execute endogenous, organized visual searches. In light, Haith confirmed Salapatek’s externality observations, formulating “visual rules” that describe how infants systematically seek out edges and track boundaries to optimize cortical firing.
• Daphne Maurer: In a series of influential studies, Maurer and Salapatek (1976) mapped infant scanpaths across real versus schematic faces in greater detail. Maurer confirmed that the shift from perimeter to internal features consistently occurs between one and two months of age, providing additional evidence for the primacy of the eye region.
• I.W.R. Bushnell: Bushnell confirmed the general timeline of internal feature penetration, but documented subtle exceptions. He showed that under optimal lighting conditions, neonates only a few days old can sometimes detect high-contrast internal elements, such as the eyes, if the stimulus is scaled and positioned carefully relative to the infant’s foveal field.

10.2 High-Density Infrared Eyetracking in Modern Developmental Science

The manual frame-by-frame photographic film analysis developed by Salapatek has been replaced by modern automated high-density infrared eye-tracking technology. Contemporary systems use high-speed digital sensors, binocular tracking, and automated computer-vision algorithms to sample gaze coordinates at frequencies up to 500 Hz or higher, without requiring physical stabilization restraints.

These modern platforms validate Salapatek’s original spatial maps while adding new analytical capabilities:

• Automated Area of Interest (AOI) analytics and two-dimensional heat maps yield fixation distributions that closely mirror Salapatek’s coordinate plots from 1975.
• Modern pupillometry captures micro-scale changes in pupil dilation, serving as a real-time measure of cognitive load, surprise, and mental effort during visual tasks.
• High-frequency tracking allows researchers to analyze micro-saccades, fixational drift, and smooth pursuit dynamics, providing deeper insights into how the preverbal infant processes dynamic, moving facial displays.

10.3 Cross-Cultural and Individual Diversity in Early Visual Scanning

A notable limitation of early infant psychophysics was its reliance on relatively small, homogeneous participant cohorts, primarily drawn from Western, educated, and middle-class populations. In recent decades, contemporary developmental researchers have expanded Salapatek’s paradigms to investigate visual scanning across diverse cultural contexts.

Cross-cultural eye-tracking studies have identified subtle cultural differences in early visual exploration:

• While infants universally prioritize the human face, their micro-scanning strategies can diverge over the first year of life based on their visual and social environment.
• Infants raised in Western contexts typically develop eye-centric, triangular scanning patterns that alternate rapidly between the eyes and the mouth.
• Infants raised in some East Asian cultural contexts often develop a more central, nose-centric fixation strategy, taking in both eyes and the mouth simultaneously via peripheral vision to prioritize social harmony and polite gaze avoidance.

These cross-cultural findings show that while the biological transition from edge-detection to internal feature scanning is a universal human milestone, the specific distribution of gaze coordinates is continuously fine-tuned by the infant’s socio-cultural environment.

11. Implications for Social-Cognitive Development, Attachment, and Atypical Trajectories

Salapatek’s work on visual scanning mechanics provided valuable insights for broader areas of child development, clinical pediatrics, and social psychiatry. The emergence of internal facial scanning is now recognized as a key behavioral building block for early social engagement, affective communication, and reciprocal social interaction.

11.1 Face Scanning as the Substrate for Intersubjectivity and Attachment

The functional shift from edge-bound to internal feature scanning at two months corresponds with the emergence of what developmental psychologists call primary intersubjectivity. This phase is characterized by reciprocal, face-to-face social interactions between infant and caregiver, mediated by mutual gaze, social vocalization, and contingent affective responsivity.

The emergence of internal facial scanning plays a direct role in supporting these social dynamics:

• By directing their gaze to the caregiver’s eyes, the two-month-old infant establishes direct eye contact, which acts as a primary release mechanism for parental caregiving behaviors, emotional attunement, and communicative scaffolding.
• Scanning internal features allows the infant to begin discriminating distinct emotional expressions, such as differentiating a smile from a neutral or fearful face.
• This mutual visual engagement serves as a foundation for secure infant-caregiver attachment, showing how basic ocular-motor maturation underpins complex socio-emotional development.

11.2 Early Biomarkers: Atypical Visual Scanning in Autism Spectrum Disorder (ASD)

Perhaps the most significant clinical application of Salapatek’s work is its direct lineage to the identification of early biomarkers for neurodevelopmental conditions, particularly Autism Spectrum Disorder (ASD). In groundbreaking longitudinal work, researchers Ami Klin and Warren Jones utilized high-density eye tracking to follow infants from birth through three years of age, mapping trajectories in infants at high and low familial risk for autism.

Their research revealed distinct visual scanning differences in early development:

• Infants who later received a diagnosis of ASD exhibited typical fixation on the eye region at two months of age, mirroring the normal transition away from the externality effect.
• However, between two and twenty-four months of age, these infants displayed a steady, progressive decline in eye fixation, accompanied by increased attention to the mouth region or surrounding physical objects.
• This atypical scanning pattern reflects a disruption in the consolidation of cortical social networks, impairing the infant’s ability to extract communicative meaning from the eyes.

Salapatek’s original methodologies and coordinate mapping techniques thus established the conceptual framework that enables modern clinical researchers to detect the earliest behavioral signs of autism in preverbal infants.

11.3 Visual Scanning Abnormalities in Preterm and Neurodevelopmentally At-Risk Cohorts

The analysis of ocular motor dynamics has also emerged as a non-invasive diagnostic probe for assessing broader neurological health in medically vulnerable and preterm infants. Infants born very preterm, or those who suffer perinatal hypoxic-ischemic insults, often exhibit significant delays in transitioning past the externality effect:

• Even after adjusting for gestational age, these at-risk infants frequently show prolonged “sticky fixations,” high fixational dispersion, and an inability to smoothly disengage their gaze from peripheral high-contrast borders.
• These atypical patterns serve as early behavioral indicators of functional compromise within descending corticocollicular circuits and thalamocortical radiations.
• Identifying these visual processing delays early has supported the creation of targeted clinical intervention protocols, utilizing gaze-directed visual exercises and structured environmental enrichment to support healthy visual and cognitive development in vulnerable populations.

12. Salapatek’s Legacy and Enduring Impact on Developmental Cognitive Neuroscience

Philip Salapatek’s career, cut short by his premature death in 1989, left an enduring mark on the landscape of modern psychology. His empirical rigor, methodological innovations, and theoretical contributions helped dismantle outmoded models of early human infancy, providing the foundation for contemporary developmental cognitive neuroscience.

12.1 Epistemological Shift: Conceptualizing the Infant as an Active Information Gatherer

Salapatek’s foundational contribution was his decisive role in dismantling the historical concept of the neonate as a passive, helpless sensory recipient. By proving that infant visual behavior is governed by precise, endogenous scanning rules, Salapatek helped replace the model of the tabula rasa with the concept of the infant as an active, computational information gatherer. His experiments demonstrated that even the youngest infants do not simply absorb light; they actively search for, isolate, and extract structural information from their visual environments.

This work established functional continuity between preverbal infant perception and adult visual cognition. Saccadic motor planning, foveal targeting, contrast sensitivity filtering, and visual disengagement are not late-emerging capacities that appear through verbal instruction; they are biologically organized systems present from the very beginning of human ontogeny. In establishing this developmental continuity, Salapatek helped integrate infant psychophysics into mainstream cognitive science and neurobiology.

12.2 Technological and Theoretical Lineage in Contemporary Research

A direct line connects Salapatek’s custom-built optical chambers from the late 1960s to contemporary tools in cognitive neuroscience. Today’s lightweight, wearable eye-tracking glasses—which record the gaze of freely moving infants in natural home settings—trace their roots directly to Salapatek’s early corneal reflection techniques. Furthermore, his early psychophysical models directly informed modern computational models of visual saliency, predictive coding, and neural network simulations of sensory filtering.

In contemporary neuroimaging, high-density eye tracking is frequently integrated with functional near-infrared spectroscopy (fNIRS) and high-density electroencephalography (EEG) event-related potentials (ERPs). This multimodal approach allows researchers to link micro-scale visual saccades to real-time hemodynamic activations in the infant temporal, parietal, and frontal cortices. Salapatek’s structural analyses also anticipate developments in computer vision and artificial intelligence, where deep learning architectures rely on hierarchical, multi-scale feature extraction strategies that mirror the biological progression from early edge detection to holistic facial categorization.

12.3 Synthesis: The Visual Scanning of Faces as a Window into the Developing Mind

Philip Salapatek’s 1975 monograph, “Visual Scanning of Faces in Infants,” remains a cornerstone text in the study of perceptual and cognitive development. His work chronicled the critical early progression of infant vision: the transition from an edge-detection system governed by subcortical reflexes into a flexible, cortically driven perceptual program capable of holistic social engagement.

Salapatek showed that human visual exploration reflects the intersection of biomechanical constraints, neuroanatomical maturation, and social ecology. Looking at a face is never a simple sensory event; it is an active, exploratory act through which an infant makes sense of their physical and interpersonal world. By charting this early developmental journey through the precise measurement of infant eye movements, Salapatek unlocked a deeper understanding of the preverbal mind, leaving an intellectual legacy that continues to guide developmental cognitive neuroscience.

Conclusion

Philip Salapatek’s pioneering investigations into the visual scanning of faces transformed our understanding of the early human mind. By engineering the infrared corneal reflection technique for use with preverbal populations, Salapatek breached the methodological barrier that had historically isolated the infant from rigorous psychophysical study. His systematic identification of the externality effect in one-month-old infants, coupled with his empirical documentation of the transition to internal feature scanning by two months of age, replaced speculative models of sensory passivity with an accurate, neurodevelopmentally grounded framework of early perception.

Salapatek revealed that the human infant’s visual journey begins not as a formless confusion, but as an organized sequence of biological milestones. As the nervous system matures, control shifts from subcortical, edge-detecting circuits to cortical networks capable of resolving the fine internal features of the human face. This developmental shift from outer boundaries to the eyes and mouth establishes the foundational architecture for mutual gaze, emotional attunement, primary intersubjectivity, and secure social attachment. Decades later, Salapatek’s micro-behavioral methodologies continue to inform modern developmental cognitive neuroscience, driving innovations in automated eye tracking, computational vision, and the early detection of neurodevelopmental conditions such as autism spectrum disorder. Ultimately, his legacy endures because he recognized that every visual saccade of the young infant represents an active, organized inquiry into the nature of the social and physical world.

References

• Bushnell, I. W. R. (1979). Modification of the externality effect in young infants. Journal of Experimental Child Psychology, 28(2), 211–229. https://doi.org/10.1016/0022-0965(79)90087-9
• Fantz, R. L. (1961). The origin of form perception. Scientific American, 204(5), 66–72. https://doi.org/10.1038/scientificamerican0561-66
• Gibson, J. J. (1979). The ecological approach to visual perception. Houghton Mifflin.
• Haith, M. M. (1980). Rules that babies look by: The organization of a newborn’s visual exploration. Lawrence Erlbaum Associates.
• James, W. (1890). The principles of psychology. Henry Holt and Company.
• Johnson, M. H., Dziurawiec, S., Ellis, H., & Morton, J. (1991). Newborns’ preferential tracking of face-like stimuli and its subsequent decline. Cognition, 40(1-2), 1–19. https://doi.org/10.1016/0010-0277(91)90045-6
• Jones, W., & Klin, A. (2013). Attention to eyes is present but in decline in 2–6-month-old infants later diagnosed with autism. Nature, 504(7480), 427–431. https://doi.org/10.1038/nature12715
• Kanwisher, N., McDermott, J., & Chun, M. M. (1997). The fusiform face area: A module in human extrastriate cortex specialized for face perception. The Journal of Neuroscience, 17(11), 4302–4311. https://doi.org/10.1523/JNEUROSCI.17-11-04302.1997
• Maurer, D., & Salapatek, P. (1976). Developmental changes in the scanning of faces by young infants. Child Development, 47(2), 523–527. https://doi.org/10.2307/1128813
• Morton, J., & Johnson, M. H. (1991). CONSPEC and CONLERN: A two-process theory of infant face recognition. Biology and Knowledge Revisited, 98(2), 164–181. https://doi.org/10.1037/0033-295X.98.2.164
• Salapatek, P. (1968). Visual scanning of geometric figures by the human newborn. Journal of Comparative and Physiological Psychology, 66(2), 247–258. https://doi.org/10.1037/h0026345
• Salapatek, P. (1975). Pattern perception in early infancy. In L. B. Cohen & P. Salapatek (Eds.), Infant perception: From sensation to cognition (Vol. 1, pp. 133–248). Academic Press.
• Salapatek, P., & Kessen, W. (1966). Visual scanning of triangles by the human newborn. Journal of Experimental Child Psychology, 3(2), 155–167. https://doi.org/10.1016/0022-0965(66)90035-7
• Salapatek, P., & Kessen, W. (1973). Prolonged investigation of a plane geometric shape by human infants. Journal of Experimental Child Psychology, 15(1), 22–29. https://doi.org/10.1016/0022-0965(73)90128-6