Facial recognition is extremely important for everyday life.It not only allows us to recognise people whom we know and may be close to, italso allows us to be aware of any dangers of people we do not recognise and donot know. It is also very important for human interaction and socialisationwhich is an aspect of life humanity would struggle to live without.
Humans areable to recognise and identify an almost infinite number of different faces anduse this to recognise different individuals. Faces have both invariant (age,sex and ethnicity) and changeable (mood and gaze) features (Haxby &Gobbini, 2010). Face recognition is thought to be a ‘special’ aspect of objectrecognition however there is evidence to suggest that face recognition andobject recognition use different neural systems within the brain and aretherefore different cognitive processes. Face recognition is very complicated,much more so than object recognition, because individuals need more than justthe basic shapes and their rough location in order to recognise a face. This isbecause faces are very similar in those basic terms, the eyes, nose and mouthof individuals are all in the same format on a face.
Facial recognitionrequires more specific and detailed knowledge of the features in order torecognise an individual. Furthermore, aspects such as expression or speech(mouth movement) which are changeable features, must not change an individual’sidentity and therefore face recognition must not change across thesealterations. There is strong evidence to suggest that recognition of faces andrecognition of expression are two independent cognitive processes and thereforedo not influence one another (Bruce & Young, 1986).Face recognition focuses on faces as a whole whereas object recognition takes astimulus apart. This is called the hypothesis of holistic face representation.
This hypothesis has been widely researched within psychology in order to gainfurther understanding into face recognition and its difference to objectrecognition, including its neural structure. Individuals have found to be ableto more accurately recognise parts of faces when presented in the whole objectthan identifying the same part but in isolation (Tanaka& Farah, 1993). This evidence supports the holistic facerepresentation hypothesis however there is still not enough substantialevidence to claim that face recognition is holistic. The definition of holismis unclear and undefined which make it difficult to put them into use inresearch.Facial recognition and attention to faces is thought todevelop from an early age or is innate. Newborns have a preference to facesover objects and patterns just only 30 minutes after being born (Johnson, Dziurawiec,Ellis & Morton, 1991). In addition,the neural systems of face recognition develop within the first six months oflife whilst faces and objects are being differentiated between (Nelson, 2001).
Johnson (1990) reviewed the development offace recognition and stated that infants shift from subcortical to corticalvisual processing within the first few months which provides an explanation ofthe possible neural pathways in infant face tracking. Research into the developmentof face recognition’s neural structures shows how the right fusiform face areaid double the size in 12-16 year olds than 7-11 year olds and even bigger inadults (Golarai, Grill-Spector & Reiss, 2006). This early researchinto the development of face recognition not only provides evidence of thespeciality of face recognition but also an understanding of how the braindevelops and more specifically, how the neural structures develop.
This essaywill discuss the neural structures of face recognition whilst evaluatingsupporting research. It will focus on the three core structures of facerecognition (OFA, STS and FFA) using different brain imaging methodologies toassess them.Face perception is argued to becarried out by domain-specific or domain-general mechanisms. Domain-specificityargues that different cognitive processes are controlled by specialised andspecific areas. Domain-general opposes this and argues that the structures usedin face recognition are not particular and are responsible for a variety ofcognitive processes (Kanwisher, 2000). Evidence supporting domain specificityis the face inversion effect (Valentine, 1988).
This proposes that faces areharder to recognise when inverted than when upright and also harder torecognise than when other stimuli are inverted. This effect provides reasonableamounts of evidence that face recognition could use its own neural systemhowever the effect could just be apparent because our initial perception offaces does not allow for inversion effects. Although this effect suggests thatface recognition is ‘special’, inversion has only been focused on faces and soother inversion effects could be possible for other stimuli. In addition, domain-specificityin face recognition is supported by cases of individuals suffering withprosopagnosia. Prosopagnosia is the inability to recognise familiar faces eitherdue to brain damage (acquired prosopagnosia) or a genetic disorder from birth(developmental prosopagnosia) (de Gelder &Rouw, 2000). Many cases of individuals suffering from prosopagnosia have beenused as evidence to show the difference between the neural structures of facerecognition and object recognition because of the individuals’ (mostly intact)ability to recognise objects still. Therefore supporting the idea ofdomain-specificity for face recognition.
However, there is still the factorthat the results of these individuals on the object recognition task are stilllower than normal which suggests that there aren’t completely specialisedregions for face recognition. Taking this into account, there is still evidencethat shows the areas affected in patients with prosopagnosia which aids in theunderstanding of regions that are used in face recognition. In her book Visual Agnosia, Martha Farah (2004)analysed these regions.
Interestingly, 65% of the patients had bilateral lesionsand 35% had unilateral lesions involving the occipital and temporal cortices,suggesting them to be neural structures of face recognition. A model proposedby Haxby, Hoffman and Gobbini (2000) proposes the specific core and extended neuralsystems responsible for face recognition. McNeil and Warrington (1993) arguethat prosopagnosia is face specific with their study of a case whom could stillrecognise his sheep however could not recognise human faces. These findingssuggest that human face recognition may be more complex because it can occur asa face-specific deficit.Regions within the brain activatedduring face recognition have been recognised through the use of brain imagingtechniques such as fMRI, TMS and EEG. Some methodologies are used together asthey reveal different aspects of cognitive processes and regions within thebrain. Brain imaging is vital for the understanding of cognitive processes andtheir neural structures. It indicates the parts of the brain responsible forcertain processes by looking at what areas of the brain are stimulated when thecognitive process is happening or what part of the brain is damaged and theside effects of this show what that part is responsible for e.
g. damage of theventral occipitotemporal cortex causing prosopagnosia leads to the conclusionthat this is one part of the brain responsible for face recognition (Sergent& Signoret, 1992). Although there are vast amounts of evidence supportingdomain-specificity in face recognition, it is not entirely clear that theseregions are only active for face recognition. Brain imaging highlights how someregions are activated for face recognition but also for other cognitiveprocesses such as the recognition of objects (Chao, Martin & Haxby, 1999).
Studies of prosopagnosia patients are flawed because tasks are not alwaysmatched and therefore these cannot state that individuals wouldn’t show similarimpairment on other tasks if properly matched (Gauthier, Behrmann & Tarr,1999). Furthermore, studies have found that the regions responsible for facerecognition can also be significantly activated by objects and animals (Ishai,Ungerleider, Martin, Schouten & Haxby, 1999; Kanwisher, Stanley , 1999). The core system is what is used toprocess invariant facial features whereas the extended systems processes thechangeable characteristics of faces. The core system for face recognition ismade up of the; inferior occipital gyri (also known as the occipital-face-areaor OFA), superior temporal sulcus (STS) and the lateral fusiform gyrus (alsoknown as the fusiform face area or FFA) regions.
Significant evidence has beenfound for the lateral fusiform gyrus in evoking activity for face perceptionand this has led to some people calling it the ‘fusiform face area’. Thisstudy used fMRI to show how the lateral fusiform gyrus was more 80% more activewhen identifying faces over objects such as cars and spoons, thereforeproviding valid evidence that this area is a structure of face recognition (Kanwisher, McDermott & Chun, 1997). The extended system is made up of the, intraparietal sulcus,auditory cortex, amygdala, insula, limbic system and the anterior temporalregions. Each responsible for other functions as well such as emotion andspatially directed speech perception (Haxby, Hoffman & Gobbini, 2000). Althoughthe occurrence of extended systems implies that face recognition is notdomain-specific, they are only aiding face recognition when working with thecore system. Therefore supporting domain-specificity because the core regionsare specific for facial recognition.
Further evidence supporting this is astudy by Perrett, Smith, Potter, Mistlin, Head, Milner and Jeeves (1984) whomidentified neurons in the superior temporal sulcus and inferior temporal cortexthat respond to faces in the brains of macaques however the evidence for humanbrains was not strong enough. Nevertheless, further experimentation andresearch could find significant evidence within human brains and so this shouldnot be completely discredited. One hypothesis about the core systems is thatthe FFA controls the processing of invariant face features whereas the STScontrols the processing of more dynamic features such as monitoring eye gaze orhandling lip reading. The study presenting this idea used fMRI brain imagingand found that the FFA responded to all faces, regardless of the expressionswhereas the STS only responded to the faces that showed strong emotion whenshowed images of faces for only 2 seconds (the invisible condition) (Jiang& He, 2006). This evidence portrays the different roles of the neural structuresfor face recognition and how they contribute to the process and supports Haxbyet al.’s (2000) model. A more complex hypothesis of face recognition comes fromresearch by Gauthier, Tarr, Anderson, Skudlarski and Gore (1999) whom proposedthat within-category discrimination and visual expertise cause activation inthe fusiform face area. This hypothesis supports domain-generality because itproposes that FFA is not specific to face recognition because response to otherstimuli could also occur here.
This is a weak argument because the results fromthe study still showed significantly larger amounts of activity in response tofaces than to the expert category, suggesting that domain-specificity is not arigid idea as some areas can still aid other cognitive processes. Furthermore,the properties of the expert stimuli (Cars and birds) both hold similarfeatures to faces (headlights and birds eyes) and this could explain why theFFA was activated by these (Turati, 2004).This is where the study could be improvedon, stimuli in way similar to faces could be used to identify if the is stillactivation within the FFA for expert categories. The study does put forward theargument that face recognition is only special because we are all experts at itbecause of its important use within everyday life and this is an idea thatneeds to be further researched.