Poster #47 - Decoding the time course of face processing in Williams Syndrome

• In Event: 2-163 - Poster Session 08
  In Poster Session: PS 08 Section - Developmental Disabilities

Integrative Statement

Introduction and Aims

The processing of face identity is often considered a relative strength in Williams syndrome (WS). Yet evidence from behavioural studies supports atypical face processing strategies in this group, e.g., reduced configural processing and attenuated inversion effects- typical hallmarks of face expertise (Karmiloff‐Smith, 2004). Critically, the neural markers of face processing ability in WS remain poorly understood. Few studies have tackled the question directly, and there has been a strong focus on characterising only the ‘classic’ face-related N170 component (Key et al., 2016; Nakamura et al., 2013). The current study uses state-of-the-art analysis, multivariate pattern analysis (MVPA), of EEG signal for a more sophisticated and detailed characterisation of the neural indices in WS associated with face sensitive processing (comparing responses to face vs house viewing) as well as configural face processing (upright vs inverted faces).

Study Population

We collected data from 11 adults with WS (7 females, Mean age =29.8±9.10 yrs) and a comparison group of 11 age matched adults (7 females, Mean age =30.27±8.78 yrs; p=.90).

Methods

Participants took part in an EEG task where they were asked to attentively watch a series of upright and inverted faces and houses. Six unique house and face identities with neutral face expressions were presented. Stimuli were presented centrally on a grey background followed by a black fixation cross. Participants completed 60 trials in each condition, alongside 60 catch trials in which a colourful butterfly appeared to the left or right of fixation. To maintain interest and attention, participants task was to press a labelled keyboard key indicating the butterflies’ appearance.
For the ERP analysis we considered the P100 and N170 components the O1/2 and P7/8 channels. MVPA was conducted using a linear SVM (support vector machine) for each timesample (downsampled to 250 Hz) across occipital electrodes (O1, O2, P7, P8, P3, P4, Pz, TP9, TP10).

Results

No effect of group was observed for the P100 component, however an altered face sensitive N170 was observed in WS. Specifically the commonly observed face inversion effect (characterized by larger amplitudes for inverted than upright faces) was present only in TD participants – though WS individuals showed the expected increased response for faces compared to houses. Further, WS individuals display an earlier N170 component for upright compared to inverted faces and for faces compared to houses.

MVPA analysis, however, revealed no difference in the onset of decoding of faces compared to houses for WS individuals, while peak decoding was significantly delayed (p=0.035). In contrast to the ERP findings, face orientation decoding was delayed in WS individuals relative to the comparison group (p=0.093) but was more sustained (p=0.007). This result may indicate that face processing is accomplished relatively faster in TD individuals.

The use of MVPA alongside the high temporal resolution of EEG allowed for the analysis of the dynamically evolving profile of face processing in WS. Through this novel application of MVPA we argue for a larger difference in the neural processing of faces in WS than what was previously understood.

Authors

• Ines Mares, Department of Psychological Sciences, Birkbeck College, London, UK
  Presenting Author

• Emily K. Farran, University College London
  Non-Presenting Author

• Annette Karmiloff-Smith, Birkbeck, University of London
  Non-Presenting Author

• Fraser W. Smith, 3 Department of Psychology, University of East Anglia, Norwich, UK
  Non-Presenting Author

• Louise Ewing, Department of Psychology, University of East Anglia, Norwich, UK
  Non-Presenting Author

• Marie L. Smith, Department of Psychological Sciences, Birkbeck College, London, UK
  Non-Presenting Author