Regional specialization of visual cortex in congenital blindness reveals takeover by multiple distinct top-down fronto-parietal inputs

Abstract

Recent evidence suggests that, in blindness, visual cortices become responsive to higher-cognitive information, including language and number. We hypothesize that this plasticity is mediated by takeover of visual cortex by multiple fronto-parietal networks that typically provide top-down feedback to vision. To test this hypothesis, we asked whether a) visual cortices participate in executive functions, which are also supported by fronto-parietal networks and b) whether executive functioning, numerical reasoning and linguistic processing colonize distinct "visual" regions. Congenitally blind and sighted control participants took part in three fMRI experiments. In a sentence comprehension experiment, participants either listened to spoken sentences or lists of non-words. In a math calculation experiment, participants solved math equations or processed sentences. Finally, we measured executive conflict using an auditory STROOP task. Participants judged whether the voice of a speaker was male or female. On congruent (C) trials, a female speaker said "female" or a male speaker said "male" and vice versa on incongruent (I) trials. On neutral (N) trials male and female speakers said gender-neutral words. In congenitally blind but not sighted individuals, parts of primary visual cortex (V1) were more active during incongruent than congruent trials (F(1,18)=4.63, p< 0.05; group interaction: F(1,26)=5.45, p=0.03). Different occipital regions, within and outside of V1, responded to language (sentence>nonwords; lateral and ventral occipito-temporal cortices (LO & VOT)) and numerical information (math>language; middle occipital gyrus (MOG)). Math- and language-responsive visual regions don't show a congruency effect (MOG: I&C>N F(1,18)=0.01, p=0.94, I>C F(1,18)=1.96, p=0.18; LO: I&C>N F(1,18)=0.28, p=0.60, I>C F(1,18)=0.27, p=0.61; VOT: I&C>N F(1,18)=1.46, p=0.24, I>C F(1,18)=0.02, p=0.88). We conclude that, in congenital blindness, "visual" cortex becomes sub-specialized for multiple distinct higher cognitive functions, including domain general executive conflict, language and numerical processing. We hypothesize that this plasticity is related to the intrinsic long-range top-down connectivity to the visual system from fronto-parietal networks. Meeting abstract presented at VSS 2018