Lecture 2 - Face Perception
preference tests
Present two stimuli at same time and measure how long infant looks at each stimulus o If the infant looks at one stimulus more than the other, it is inferred that they 'prefer' this stimulus o Thus, the infant can discriminate between stimuli and for some reason, finds A more stimulating than stimulus B o Whatever the infant finds particularly stimulating gives clues about which aspects of the environment are contributing to their development o Contemporary research has moved away from the researcher observing how long the infant looks at each stimulus to using eye-tracking technology o Indicates aspects of the environment the infant finds most stimulating o Fantz (1961): Bars = proportion of time spent looking in comparison to other patterns. Complex patterns are viewed for longer, but schematic faces have the highest time proportion
conditioning
Repeatedly reward target behaviour so infants perform reinforced behaviours o Experimenter must wait for the infant to naturally perform the behaviour in order to reward it o E.g. increase sucking rate, get specific stimuli like photo of caregiver o Infant becomes habituated to stimulus and stimulus is altered to a new photo (e.g. HAS procedure) If infant does not increase sucking rate treats 2 stimuli as the same Does increase sucking rate distinguishes between 2 stimuli
habituation tests
o Infant is repeatedly shown a stimulus and it is likely each time it's presented the infant will spend less time looking at it. If the stimulus is changed, infant shows a renewed interest in the novel stimulus o Habituation = Shown interesting stimulus repeatedly and Infant loses interest eventually o Dishabituation = Change to a different stimulus - Infant shows renewed interest and looks again o If so, the infant can tell the difference and distinguish the stimuli
face perception
• "Faces are arguably the most important visual stimulus used in human social communication" Moulson et al (2009), p.1 • Why is it useful? What abilities are we born with? • How does it develop? Is it a specialized ability? When is it fully mature?
abnormal face perception
• ASD deficits in social cognition: o Recognizing familiar people and remembering faces o Interpreting eye-gaze and emotions • William's Syndrome o Process unfamiliar faces atypically o More interested/ intuited to faces o Rate faces as more approachable/ friendly than controls o Prolonged face gaze o (Riby et al. 2008) • Prosopagnosia (face blindness) o Damage or abnormalities in right fusiform gyrus (stroke, brain injury) o Congenital prosopagnosia - from birth, appears to run in families o Different degrees of severity: might not even recognize own face
beyond infancy: continuing development
• Adults are experts o Recognize face as familiar within 0.5s o Retain information of large number of faces o 90% recognition of yearbook photos from Class size of up to 900, up to 35 years later • So, if adults are experts, when does this expertise fully emerge? o Some research suggests not until 30+ years for face learning/recognition o Why does it take so long?
Sugita (2009) Innate Face Processing
• Before seeing any faces or face-like objects, macaque monkeys showed a preference for faces rather than nonface objects. They showed remarkable face processing abilities both for human and monkey faces • Macaque new-borns could imitate human facial gestures, indicating the ability to match their own facial movements to observed facial gestures • It seems likely that new-borns acquire the knowledge about the basic structure of their own face, presumably through proprioception, so that facial structure would become a familiar and attractive visual object without the experience of the face itself
visual perception development
• Develops very rapidly from birth and primarily in first year • Visual acuity = ability to distinguish visual detail and definition o Poor at birth, rapid increase in first 6 months o Near adult levels by 1 year o Newborn infants can only detect the separation of the lines if they are about 30 times wider than the minimum width adults can detect o E.g. wouldn't be until the later paddles that young infants would distinguish the stripes as separate lines • Visual scanning = ability to track moving objects across the visual field o < 2 months, cannot track moving objects smoothly, instead they follow moving objects with a series of jerky eye movements o Salapatek (1975) investigated eye movements of young infants as they scanned geometric shapes. o 1 month: focus on limited features of shape, particularly outside edges o 2 months: start to focus on internal features • Colour vision o Newborns have a limited ability to distinguish colour: can distinguish between white and red, but not other colours (e.g. Adams et al. 1994) o Around 1 month, look longer at brighter, bold colours o By 4 months, close to adult ability to discriminate colours • Young infants have a functional and effective visual system, but the quality of their vision is poorer than adults • Infants have early preferences for complex patterns o Fantz (1975) 1-month-old infants preferred the less complex images compared to 2-month-old infants who preferred the more complex stimuli
effect of environment
• Environment effects the type of faces that we are faster at recognising • Narrowing of the "perceptual window" o As we get older, face-perception skills become more specialized in the type of faces we are more likely to see o Pruning of synapses • Pascalis et al (2002) Is face-processing species-specific during the first year of life? o Tested young children's ability to distinguish between monkey faces and human faces using a visual paired comparison (VPC) task o VPC task - recognition memory task that assesses the proportion of time an individual spends viewing a new picture compared to a picture they have previously seen i.e. novelty preference Image of face A was projected, then there was a blank screen for and the two images (one familiar, face A, and one novel) were presented side by side Alternated trials with using 2 human stimuli and 2 monkey stimuli o Tested participants aged 6 months, 9 months and adults (mean age 28.5 years) o If the youngest infants could discriminate between the human faces, it could have been because they have some very early ability to do so or because even 6-month-olds have a great experience of human faces o The youngest infants would have very little experience of monkeys and if they could distinguish between two monkey faces, it would suggest a general ability to discriminate between all types of faces o Results For adults and 9 months, the mean time spent looking at the human novel face was significantly longer than spent on the familiar face. The preference for the novel face demonstrated that they could distinguish the two human faces However, adults and 9-month-old infants looked as long at the novel and familiar monkey face, implying the adults could not distinguish the two monkey faces 6-month-old infants preferred to look at the novel human face and the novel monkey face, hence 6m infants could discriminate between different monkey faces and different human faces 9m infants and adults could only discriminate between human faces If exposed to monkey faces (parents reading books about monkeys every day for 2 weeks), 9ms could discriminate between 2 different monkeys but not at the same level as 6m infants o Perceptual window for face processing narrows during the first year as very young children can distinguish faces of another species, older children and adults don't have this ability o Face-perception skills specialize for just faces of one's own species o Our environment impacts face recognition of our own species • "Other-race" effect o Adults are poorer at discriminating faces of other races compared to own race (Tanaka et al., 2004) o Driven by the environment as we tend to be exposed to certain races o 3m old, but not newborns, prefer own race faces (Kelly et al 2005) • Sangrigoli et al (2005): Korean adults adopted between 3-9 months into Caucasian families were more accurate with Caucasian faces o No genetic predisposition to the type of faces we perceive • Early social experience o Better at discriminating & recognizing female faces Not universal so perhaps primed by caregivers who are usually female Effect of exposure to primary caregiver Preference for female faces in 3-month-old, not newborn, infants (ex: Quinn et al. 2008) Fathers as primary caregivers = preference for male faces (Quinn et al. 2002) o Institutionalized children showed deficits in identifying emotions in faces (Wismer Fries & Pollak, 2004) - lack of social input o Children raised in abusive environment show bias for angry faces (Pollak et al. 2000) - this is the stimuli most prominent to them
Fantz (1961) The Origin of Form Perception
• Experiments indicate that the ability to perceive the form of objects is innate but that maturation and learning play important roles in its development • Argued that very young infants cannot respond to stimuli as shape, pattern, size or solidity and thus, cannot perceive form • 1 to 15-day old infants showed a preference for roundness, spherical objects compared to triangular 3D objects showing some degree of form perception is innate • There is an adaptive significance of form perception - infants preferred a 'real' face over a scrambled face and the control light/ dark face
Pascalis (2011) Development of Face Processing
• Faces can be organised into face traits and face states • Face traits = refer to the visual information in the face that is relatively permanent and stable e.g. gender, race, age, species • Face states = dynamic and transient facial cues used to process speech, emotional expressions, attention and intentions • Viewing a face triggers at least two automatic and fast processes: o Categorization of the stimulus as a face belonging or not belonging to our own group or species o Recognition of the face at an individual level • Valentine (1991) suggested that faces are encoded in a face-space framework in terms of vectors in a multidimensional perceptual space o The origin of space represents the average of all faces experienced by an individual o More typical faces are closer to the origin and more distinctive faces are further out in space o New faces are encoded in terms of their deviation from this norm face o By age 5, the face space of young children approximates the face space demonstrated by young adults o Face-space model explains the other-race effect: face spaces are tuned toward the faces present in the environment • Infants process faces as members of various categories, a manner similar to adults o During the initial months of life, they categorize stimuli into human versus nonhuman categories, cats versus dogs, and even different breeds of cats and dogs o In the next several months, they categorize human faces into male versus female and prefer female faces, attractive versus unattractive faces and prefer attractive faces, and Caucasian versus Chinese faces. o They also categorize faces in different age categories, preferring child over adult faces. o In addition, Rennels and Davis (2008) reported that infants experience far more female faces than male faces in their daily interactions with others. Quinn et al. found that infants who were raised by mothers preferred female over male faces while those raised by fathers preferred male over female faces. o Kelly et al. showed that both Chinese and Caucasian 3-month-old infants prefer looking at their own-race faces than at other-race ones. o These findings suggest a role for experience in early face category processing. Children are not simply processing individual faces; they also categorize faces according to race, species, gender, and age, and show preferences for one face category over another that are due to selective experience. • Configural = the emergent features (eyes, ears, mouth, nose) of a face when two or more features are processed at the same time. o E.g. when visually processing both the eyes and mouth of a face, this would be configural processing • Encoding switch hypothesis posited that before 10 years of age, children relied mainly on isolated facial features to recognize faces and after 10 years children begin to rely on configural information o However, infants as young as 3 months and children below 10 have been found to be sensitive to configural information and use this information in face identity processing o E.g. sensitivity to configural changes among facial features emerges between 3 and 5 months of age o Slater (2000) found that new-borns' preference for attractive faces is disrupted when the faces are inverted o Quinn (2009) found that 3-month-old infants' preference for female faces over male faces is not found if the faces are presented with inversion o Many studies have demonstrated the existence of inversion effects at a young age and these findings refute the encoding switch hypothesis • Young children may have more difficulty processing configural than featural information e.g. Mondloch (2002) showed that children do not reach the level of adult performance in a configural task until 15 years of age o Whereas younger children show greater sensitivity to featural information relative to configural information, their configural processing ability still undergoes significant development from childhood to adolescence • Infants present a preference for face stimuli and facial discrimination using featural, configural and holistic cues o This early competence is later refined as evidenced by age-related changes throughout childhood o Some refinements are due to the development of general cognitive abilities where as some (e.g. configural processing) may be face specific
Innate face preference
• Fantz (1961): 1-15-week-old prefer complex patterns (schematic and jumbled face) o Very young infants can discriminate between patterned and unpatterned shapes o E.g. they preferred striped or checkerboard patterns to plain discs o Infants view the complex patterns equally, but prefer it over the simplistic pattern of a light-dark difference • Maurer and Barrera (1981): add controls for complexity with 3 equally complex stimuli o One stimulus is a schematic face, one a symmetrical scrambled faces and the 3rd an aysmmetrical scrambled face o All had the same facial features, and thus, same complexity o 1 month: no difference in looking times for the 3 stimuli o 2 months: looked longer at "natural face" o Any preference for natural faces at 2 months was more than just a preference for complex stimuli (Previous preference in new-borns was for complexity) • Goren et al (1975) o Used moving stimuli instead of static with newborns o Showed a schematic face, a schematic symmetrical scrambled face or a blank outline o Newborns tracked schematic face more than other two, suggests infants have some ability to detect face-like stimuli from birth
Crookes and McKone (2009) Early Maturity of Face Recognition: No Childhood Development of Holistic Processing, Novel Face Encoding or Face-Space
• Full quantitative maturity is reached early, by 5 - 7 years at the latest, perhaps earlier • Face-specific perceptual development theory = an important contributing factor is ongoing development of face-specific perceptual mechanisms (e.g. holistic processing, tuning of face-space dimensions) o Face perception continues to develop into late childhood, due to experience with faces o Ongoing improvements in face coding contribute directly to improvements on perceptual tasks like face discrimination o Improvements might occur in holistic/ configural processing o Strong perceptual integration of information across the whole face and processing of the 'face-space' o Dimensions of face-space are determined through experience, and tuning continues through life. It has been proposed that children might use fewer dimensions than adults, or the same dimensions but differently weighted, or might code discriminations along each dimension less finely or that the occupation of children's face-space by fewer familiar exemplars might functionally affect face perception (Humphreys & Johnson, 2007) o Development in the ability to perceptually encode a novel face as young children don't form representations of newly encountered faces as efficiently as do adults o Face-specific perceptual development theory argues the improvement seen on face tasks between 5 years and adulthood results substantially from changes within the face perception system • General cognitive development theory = face perception itself is mature in early childhood, and that all development of test performance thereafter reflects improvements in general cognitive mechanisms o Improvement seen on face task after early age is due entirely to the development of general cognitive factors e.g. concentration, visual attention which are known to improve with age o Perceptual coding of faces is fully mature early so all subsequent development on task performance can be explained by development of other factors
genetic differences
• Inherited predisposition in facial recognition accounts for individual differences in face perception • Twin studies reveal strong genetic influence • However, depends on the face o Differences between familiar and unfamiliar faces (Young, 2018) o Humans are between at recognising familiar faces • Task differences: poor at recognising faces from different angles o E.g. CCTV footage
early face preference
• Johnson et al (1991) Replicated Goren's effect with newborns o By 3 months, no longer track schematic face anymore than the scrambled and blank face o Schematic face is no more exciting than the other 2 paddles o Believed early face recognition is based on 2 processing systems for this face preference to vanish • Johnson and Morton (1991) 2 process model: o Conspec = Early system (subcortical structures) biases infants to orient towards faces during the first weeks of life to draw infant's attention to moving faces o By attending to moving faces, infants can learn about different faces o Conlearn = Later taken over by more mature system (visual cortex) around 3 months with more precise recognition that allows infants to distinguish human faces and other stimuli, as well as static faces o Conlearn takes over around 3 months; at this point infants are more interested in subtle features and details of a face, so the schematic face is less interesting o Exposure of faces over the first few months leads to conlearn, i.e. develop interest in detail of faces • What else can new-borns do? o Recognize identity of novel individuals (Turati et al., 2008) o Recognize eye-gaze: Look more at direct than averted gaze (Farroni et al., 2002) o Recognize expressions: Infants dishabituated when expression changed (Field et al., 1982) o Prefer attractive faces: Newborns < 1week old looked longer at attractive faces (Slater et al. 2000) • Newborns can discriminate their mother's face o E.g. Walton et al (1992) infants sucked more to keep mother's face on video at 1-4 days old o How are they doing it? o Pascalis et al. (1995) 4-day-old newborns could discriminate between their mother and an unfamiliar female However, preference for mother's face disappeared when outside of face and hairline masked Newborns use outer features to identify (middle row), infants track outer features and external contours before they develop into noticing inner features If newborns don't process inner features, we would not expect them to recognize static faces o Turati et al. (2006) Could use both outer and inner features (middle and lower row) • Infants show very early preferences for faces, and even certain types of faces o Discriminate between different faces and even features • Suggestive, but not conclusive, that face perception must be innate to some extent o Infants have heavy exposure to faces from birth, so difficult to disentangle nature/ nurture • Sugita et al (2008) Monkeys not exposed to faces for first months of life still preferred them o Suggestive of innate face perception
theoretical approaches
• Nature vs. Nurture o Nativism: abilities from birth - innate, inborn o Empiricism: acquire overtime through experience - learned • Empiricists (John Locke) argued that the newborn infant is a tabula rasa (blank slate) on which experiences are imprinted o E.g. William James (1890) stated that sensory inputs become fused into 'one blooming, buzzing confusion' and it's only through experience that children can discriminate among them • Nativists (Descartes, Kant) argued that perceptual abilities are present at birth and innate o Gestalt psychologists stated that certain perceptual abilities were present at birth because of the structural characteristics of the nervous system and that the infant actively tries to create order/ organize their perceptual world • Researchers have found that infants are born with a wider range of perceptual abilities than empiricists suggest and that infants' capacity to learn rapidly from experience is greater than nativists propose • Are faces special? o Special perceptual process: organized at birth o Perceive faces as they perceive other objects: becomes specialized after experience
Perception
• Sensation = information about environment picked up by sensory receptors and transmitted to brain • Perception = interpretation by the brain of sensory input o How we understand the events, objects and people in our environment
late maturation vs early maturation
• Two key theories: • Face specific perceptual development theory: o Ongoing development of face-specific perception mechanisms continue to develop into late child and adolescence (late maturation) o Face perception gets better because of increased exposure/experience with faces • General cognitive development theory: o Face perception matures early (4-5 yrs.) (early maturation) o Adults are better than younger adults as performance increases later as general cognitive mechanisms improve o See: McKone et al. 2012; Crookes & McKone, 2009 • When does it mature? Early research suggests qualitative change later in childhood/adolescence • Adult mechanisms of face perception? o Disproportionate inversion effect More accurate when faces are upright Upside-down faces are very difficult for us to perceive Larger effect for face versus non-face objects (e.g. a chair) o Holistic/configural processing Integration of information from all regions of face E.g. distance of eyes from the nose Code spacing between face and features o More recent research suggests adult-like mechanisms in place much earlier (Crookes and McKone, 2009) As young as 4 - 5 years Suggests that increase reflect development of other cognitive abilities (concentration, attention, memory) o Susilo et al. (2013) Tested over 2,000 18-33-year olds, controlled for non-face visual recognition, sex, and own-race bias Positive association between age and facial recognition abilities, not for abstract art Facial recognition continues to develop over time, whilst e.g. non-face visual recognition does not Conclude results support "late maturation hypothesis", early maturation hypothesis would assume no differences in face recognition between 18 and 33-year-olds
why is face perception useful?
• What can you tell from a face? o Species o Sex o Race o Identity o Mood, Emotional state o Intent, truthfulness • Crucial ability for successful social life
Turato and Cassia (2006) Newborns' Face Recognition: Role of Inner and Outer Facial Features
• With development, the perceptual cues that children use to recognise people's faces change • Specifically, the relative importance of inner and outer features of the face changes with age in relation to the familiarity of the person to be recognised • For high familiar faces, there is a transition from greater reliance of outer features to inner features (Campbell & Tuck, 1995) 5-to-6-year-old children are more likely to be able to name their classmate or famous people from seeing just their outer features. At 11, there is a reversal in this pattern of reliance on outer over inner parts for recognition and as adults, they rely more on inner facial features. For unfamiliar faces, in all age's recognition is easier when driven by outer facial cues • Thus, regardless of the familiarity, outer facial cues are the perceptual cues behind children's facial recognition • Between 3 and 10 months, infants move from processing the internal and external features of a face independently to processing the relationship between them (Cashon & Cohen, 2003). Shows the developmental changes in infants' capacity to process faces in a configural rather than analytical matter
