Here are the analysis codes for the "all-or-none neural mechanisms underlying face categorisation: evidence from the N170" article, which has been accepted in Cerebral Cortex. The preprint version can be found: https://psyarxiv.com/qu47m/. The accepted manuscript can be downloaded from my website. This study is also part of my PhD thesis (Chapter 5).
Categorisation of visual stimuli is an intrinsic aspect of human perception. Whether the cortical mechanisms underlying categorisation operate in an all-or-none or graded fashion remains unclear. In this study, we addressed this issue in the context of the face-specific N170. Specifically, we investigated whether N170 amplitudes grade with the amount of face information available in an image, or a full response is generated whenever a face is perceived. We employed linear mixed-effects modelling to inspect the dependency of N170 amplitudes on stimulus properties and duration, and their relationships to participants’ subjective perception. Consistent with previous studies, we found a stronger N170 evoked by faces presented for longer durations regardless of subjective confidence in perceptual categorisation. However, further analysis with equivalence tests revealed that this duration effect was eliminated when only faces perceived with high confidence were considered. Therefore, previous evidence supporting the graded hypothesis is more likely to be an artefact of mixing heterogeneous “all” and “none” trial types in signal averaging. These results support the hypothesis that the N170 is generated in an all-or-none manner, and, by extension, suggest that categorisation of faces may follow a similar pattern.
The codes used for data analysis and the outputs are available here.
The codes for plotting ERP plots are available here.