Week 11 - faces

• the brain sometimes recognises faces in objects when it shouldn't or there isn't one
• age, race, sex, health etc are all available in a face as well as their identity
• the L and R fusiform gyrus ability to detect faces in organisms and nature leads to a phenomenon called pareidolia
• newborns show preferences for cartoonish faces as well as turning their eyes and head most for a face configuration
• 6 month olds can discriminate between a human and monkey face (Pascalis, 2002)
• 9 month olds can discriminate human faces bu lose the ability to discriminate between monkey faces
• features entail eyes mouth etc. Each identifiable parts that vary in subtle ways across individuals
• configuration is the arrangement of these features such as spacing, symmetry, position within the face etc
• holistic (configural) face processing - involves integrating info from an entire object
• there is evidence that faces (not other objects) are recognised through holistic processing
• configural processing is crucial for getting a likeness in a portrait, meaning it is crucial to person recognition
• configural processing means how our brains look at the overall shape and arrangement of things, rather then individual parts such as how we look at a face
• when configural processing is disrupted face recognition is hard
• experiments use composite faces separating top and bottom to show how recognition becomes dramatically impaired supporting the importance of configural processing
• details of the face such as features seem less important in recognition then the configuration as this causes the likeness largely
• when using a 50/50 face the neural code becomes biased to the one looked at longer or the one previously unviewed then causing the other to become more viewed
• adapting to a happier face then makes a neutral face looks sadder when it actual fact it is not and vice versa due to emotion after effect
• 50/50 thing is identity after effect
• coding expression and identity use separate pathways
• faimiliar and unfamiliar faces are processed differently
• Bruce and Young suggested 2 main stages for processing: Structural encoding - deals with viewpoint, lighting and visual analysis of object. Extended processing splits to two separate pathways: expression analysis and face recognition.
• Put surface electrodes on the scalp recording voltage change. Voltage measurement is started when a picture onsets. This process is repeated then average the signals in a time locked manner. The resultant waveform is the ERP.
• a strong negative going signal appears about 170ms post stimulus when faces are presented and is known as N170 and is an ERP signature for faces
• N170 is more obvious for face pictures than other objects and assumed to reflect structural encoding due to not varying with expresion, familiarity and viewpoint
• fMRI can be used to find specific areas the brain is activated by more for one category then the other
• Repetition suppression - brain activity reduced for stimulus repetition. Areas showing RS are more sensitive to the repeated feature
• different images of the same person making different expressions were presented. If RS occurs then the brain area is sensitive to face identity and not to facial expression. Found that RS is found in FFA and posterior STS. These areas see identity not expression
• prosopagnosia results in failure to overtly recognise people even if covert measures suggest recognition. Ventral route effected including FFA
• Capgras delusion results in feeling people are imposters even though they are overtly recognised. Dorsal route affected including STS
• facial expressions involve muscles near eyes and mouth - genuine smile involves eyes and mouth , polite is only mouth
• negative expressions involve the eyes, eyebrows, mouth and different muscles are used for negative compared to positive
• eye movements to faces reveal how info is acquired. Reading a face required attention to eye muscles, mouth position and their effects on nose, cheeks and chin
• some struggle reading facial expressions such as patients with amygdala damage and autistic people - particularly children. Spend more time looking at the mouth then eyes and this is the opposite to normal kids
• reduced perfomance in differentiating faces based on both sex and individual identity in autism
• no evidence for face inversion effect (reduced holistic processing). Reduced brain activation for faces
• Correlation between reduced face processing and increased ASD behaviour
• face space believes each face has a MDS which is unique with its own important features and axis
• Individuals are recognised by how a set of features deviate from the centre with the size of these deviations coded and then used to identify an individual
• Each person has their own “average face” because the MDS for each face is developed from personal experience. 
• recognition is better the further a face is from the average