Abstract
Face detection is fast, accurate, and seemingly effortless. Here we introduce a highly sensitive method to measure face detection thresholds rapidly, objectively and independently of low-level visual cues. The method is based on the swept parameter steady-state Visual Evoked Potential (SSVEP), in which a stimulus is presented at a specific ("tagged") frequency while parametrically varying ("sweeping") the detectability of the stimulus. For this experiment, the stimulus consisted of the appearance of a face image (F) from noise comprised of a phase-scrambled face image (SF) with an equal power spectrum and mean luminance. Alternations between the images (F-SF-F-SF-F-…) at a constant rate (3/second) elicited a robust odd harmonic response (3.0 Hz) specific to the structure of the face. The visibility of face images was increased by progressive de-randomization of their phase spectra in a series of equally spaced steps. Trials contained an ordered series of 20 levels of image degradation presented over 20 seconds. High-density EEG was recorded from 10 human adult observers. While face information was revealed only gradually in the stimulus, the evoked response at the first harmonic (3.0 Hz) emerged abruptly. The face-specific detection response was most prominent on right lateral occipito-temporal sites. Thresholds for face detection were estimated reliably from the emergence of the first harmonic response in single observers from 15 trials, or on each of the 15 individual face trials across observers in the study. The VEP derived thresholds correlated with the concurrently measured behavioral face detection times. This first application of the sweep VEP approach to high-level vision provides a sensitive and objective method that could be used to measure and compare visual perception thresholds for various object shapes and levels of categorization in different human populations, including infants and individuals with developmental delay.
Meeting abstract presented at VSS 2012