Naturalistic stimulation in sensory-deprived individuals reveals overlapping large-scale brain organization with differential cross-modal mechanisms

Animal and human sensory-deprived models offer the possibility to study the causal mechanisms implicated in perceptualprocessing and knowledge organization and investigate how thebrain copes with the absence of modality dependent information.Congenital blindness and deafness represent unique tools tocomprehend to what extent (the lack of) a specific sensory input isa necessary condition for the morpho-functional development ofboth early sensory and higher-level brain regions. Accumulatedevidence from animal studies and comparative works in humanswith congenital/early sensory loss has reliably described majorgeneral tenets: preservation of modality-independent large-scaleorganization with topological/regional task selectivity, and crossmodal,modality-dependent plasticity phenomena. Indeed, whilston one hand the large-scale architecture of the brain conservestask specificity, at the local level, early sensory areas deprivedsince birth consistently show cross-modal engagement for theinformation coming from spared senses (e.g. auditory and tactilestimulation for the visual cortex in congenital/early blindindividuals), whereas higher level cortical regions in both theauditory and visual hierarchy are still able to represent stimuluscharacteristics regardless from the sensory modality conveyingthe information to the brain. Although robust and replicable, thesefindings leave open questions about the functional mechanismsunderlying the observed brain reorganizations. Recently, a growing number of studies have used naturalisticstimulation in fMRI to convey complex, real-life-like perceptualand semantic information thus fulfilling the need of ecologicalvalidity and generalizability of results to daily-life perception andcognition. Indeed, contrary to traditional paradigms, that makeuse of simplified, artificial stimuli (static images, visualgeometrical patterns or isolated sounds), the setup of naturalisticexperiments consists in the usage of prolonged, complexstimulation with the aim to reflect more faithfully the dynamicstructure of natural environments humans experience on a dailybasis.To this regard, movies and books audiodescriptions have beenwidely used to convey a continuous and rich stream ofinformation whose processing requires the concerted deployment and combination of different brain mechanisms, going from theonline integration of the incoming multisensory perceptual inputto more complex cognitive operations that impinge on attention,information retrieval/update and semantic knowledge. Thus, therichness of such paradigms offers the unique possibility to studybrain functioning in complex settings that closely mimic everydaylife experiences.My PhD research meant to provide new insights on the role ofexperience-dependent plasticity in shaping brain functioning ineveryday life. The project was carried out through a 3T functionalMagnetic Resonance Imaging (fMRI) study with the aim toevaluate and compare patterns of brain response to a prolongednaturalistic stimulus (@ 50 minutes long movie) in the earlysensory areas of two models of sensory-deprivation, congenitalblindness and deafness. The film One Hundred and OneDalmatians (Walt Disney, 1996) was shortened and edited (withthe addition of subtitles and audio descriptions) in order to createthree different versions: two unimodal conditions (i.e. auditorymoviewith audiodescriptions or visual-movie complemented bysubtitles) and one multimodal setting (audio-visual). Specifically,we measured brain responses in five experimental groups ofparticipants: congenitally blind (n=9; 44±14 years), congenitallydeaf (n=9, 24±4 years), and three control groups that attended theauditory (n=10, 39±17 years), visual (n=10, 37±15 years), or audiovisual(n=10, 35±13 years) variants of the movie.After standard preprocessing, we took advantage of an Inter-Subject Correlation (ISC) analysis to measure to what extentsensory deprivation affected whole-brain participantssynchronization in each group separately. We then comparedcongenitally deaf and blind with their control groups,demonstrating wide and overlapping modality-independentresponses across groups, accompanied by a less lateralizedrecruitment (e.g., higher ISC in the right hemisphere) in both blindand deaf individuals. Afterwards, we focused our analysis onsensory areas, showing how V1 and A1 and late regions weredifferently affected in the congenitally deaf and blind groups.Finally, through computational modeling, we further described ineach brain region to what extent the ISC was dependent to specificstimulus characteristics, in terms of both low- (i.e., visual orauditory) and high-level (i.e., semantic) features. Results clearly indicated that V1 functional activity in blind individuals wasdriven by acoustic features (i.e., sound envelope), whereas the roleof A1 in deaf individuals was not related to any of the low-levelnor high-level stimulus descriptions explored. In this thesis, Ipresented the methodology and the result of the study anddiscussed the implications of the findings as compared to thecurrent evidence in the literature.