Effects of Part- and Whole-Object Primes on Early MEG Responses to Mooney Faces and Houses.

Mara Steinberg Lowe,David Poeppel,Gwyneth A Lewis

doi:10.3389/fpsyg.2016.00147

Mara Steinberg Lowe, David Poeppel + Show 1 more

Open Access

https://doi.org/10.3389/fpsyg.2016.00147

Copy DOI

Abstract

Results from neurophysiological experiments suggest that face recognition engages a sensitive mechanism that is reflected in increased amplitude and decreased latency of the MEG M170 response compared to non-face visual targets. Furthermore, whereas recognition of objects (e.g., houses) has been argued to be based on individual features (e.g., door, window), face recognition may depend more on holistic information. Here we analyzed priming effects of component and holistic primes on 20 participants' early MEG responses to two-tone (Mooney) images to determine whether face recognition in this context engages “featural” or “configural” processing. Although visually underspecified, the Mooney images in this study elicited M170 responses that replicate the typical face vs. house effect. However, we found a distinction between holistic vs. component primes that modulated this effect dependent upon compatibility (match) between the prime and target. The facilitatory effect of holistic faces and houses for Mooney faces and houses, respectively, suggests that both Mooney face and house recognition—both low spatial frequency stimuli—are based on holistic information.

Highlights

Stimulus selectivity is a major focus in studies of face and object perception
The amplitude was significantly greater for faces than houses in both the M100 [t(17) = 2.28, p = 0.036, d = 0.38], and M170 responses [t(17) = 2.77, p = 0.013, d = 0.64]
A number of MEG and EEG studies have reported enhanced M100 or P1 amplitude for faces presented as photographs

Summary

Introduction

Stimulus selectivity is a major focus in studies of face and object perception. While object recognition has been argued to be based on featural representations (e.g., doors and windows represent houses; cf Tanaka and Farah, 1993), face recognition may involve holistic/configural information (Maurer et al, 2002). “Featural” refers to individual components (e.g., nose size), while “configural” refers to relations between components. Maurer et al (2002) distinguish configural processes as first-order relational—(structure/arrangement of features, e.g., eyes above mouth), holistic—(merger of features, e.g., a gestalt), and second-order relational—(spacing between features). Schematic faces (e.g., smiley faces) lacking meaningful physiognomic information should not engage second-order processing (Sagiv and Bentin, 2001) whereas two-tone Mooney faces lacking first-order features that are individually recognizable must be recognized holistically (Latinus and Taylor, 2005)

Methods

Results

Conclusion