The goal of this proposal is to better characterize the cognitive and neural basis of face perception heterogeneity in developmental prosopagnosia (DP) and to test whether potential perceptual subtypes are better suited for one type of cognitive training program vs. another. This is relevant to the National Eye Institute's mission to better treat visual disorders and understand mechanisms of visual function. Our pilot results (N=45 DPs) provide preliminary evidence that there are perceptually-impaired (PI-DPs) and perceptually-unimpaired (PU-DPs) DP subgroups that may be mechanistically distinct. We find that PI-DPs have deficient eye processing, reduced holistic processing abilities, reduced N170 response to eye contrast- reversed faces, and a trend towards decreased white matter integrity in the occipital face area, a face-selective region involved in face parts processing. We also find that PI-DPs, compared to PU-DPs, show a greater treatment response to our computerized face perception training program. The goals of the current proposal are to build on these pilot results to further test and validate the PI-DP vs. PU-DP subgroup distinction, to better understand the neural mechanisms underlying perceptual heterogeneity in DPs, and to examine treatment implications of potential subgroups. In particular, our aims for this proposal are to: 1) Collect a large sample of web-based and in-lab DPs (N=280) and controls (N=140) and replicate our pilot findings showing eye processing and holistic processing deficits in PI-DPs. We will also perform latent profile analysis to determine if DPs naturally form PI-DP vs. PU-DP subgroups or rather represent a continuum of face perception deficits. 2) Characterize the neural mechanisms of DP face perception heterogeneity using functional fMRI, EEG, and diffusion tensor imaging in 80 DPs and 40 controls. To test better understand the neural underpinnings of DPs' reduced eye region sensitivity and reduced holistic processing, we seek examine the N170 event-related potential response to isolated eyes vs. mouths as well as face inversion and for fMRI, population receptive fields of face-selective regions as well as responses to isolated mouths and eyes. We will also examine whether PI-DPs vs. PU-DPs have white matter integrity differences in face-selective white matter regions. 3) Determine whether PI-DPs vs. PU-DPs have differential responses to face perception training, face memory training, or face perception+face memory training programs. This will involve a longitudinal study of 120 DPs being assigned to either a waitlist control condition or 6 weeks of one of the three cognitive training programs. Before and after training/waiting, we will assess DPs on a validated battery of face perception and recognition tests as well as self-reported face recognition. To measure the longevity of potential training effects, DPs will also repeat assessments after a 6-week no-contact period.