A multimodal neural signature of face processing in autism within the fusiform gyrus

Background Differences in face processing are commonly reported in case/control studies of autism. Their neural correlates have been explored extensively across single neuroimaging modalities within key regions of the face processing network, such as the fusiform gyrus (FFG). Nonetheless, it is poorly understood how different variation(s) in brain anatomy and function combine to impact face processing and social functioning. Extracting the shared information across different modalities is essential to better delineate the complex relationship between brain structure and function, leading to a more comprehensive understanding of the mechanisms underlying autism. Methods Here, we leveraged data from the large multimodal EU-AIMS Longitudinal European Autism Project (LEAP) to study the cross-modal signature of face processing within the FFG across structural magnetic resonance imaging (MRI), resting-state fMRI (rs-fMRI), task-fMRI (based on the Hariri emotional faces task) and electroencephalography (EEG; recorded when observing facial stimuli) in a sample of 99 autistic and 105 non-autistic individuals (NAI) aged 6-30 years. We combined two methodological innovations: (i) normative modelling was employed on each imaging modality separately to derive individual-level deviations from a predicted developmental trajectory and (ii) unimodal deviations were fused through Linked Independent Component (IC) Analysis to simultaneously decompose the imaging data into underlying modes that characterise multi-modal signatures across the cohort. Next, we tested whether ICs significantly differed between autistic and NAI and whether multimodal ICs would outperform unimodal ICs in discriminating autistic individuals from NAI using a support vector machine under 10-fold cross-validation. Finally, we tested the association between multimodal ICs and cognitive, clinical measures of social or non-social functioning in autism using canonical correlation analysis (CCA). Results In total, 50 independent components were derived. Among these one multimodal IC differed significantly between autistic and NAI ( t =3.5, p FDR=0.03). This IC was mostly driven by bilateral rs-fMRI, bilateral structure, right task-fMRI, and left EEG loadings and implicated both face-selective and retinotopic regions of the FFG. Furthermore, multimodal ICs performed significantly better at differentiating autistic from NAI than unimodal ICs ( p <0.001). Finally, there was a significant multivariate association between multimodal ICs and a set of cognitive and clinical features associated with social functioning ( r =0.65, p FDR=0.008); but not with non-social features. Discussion The FFG appears to be a central region differentially implicated in autistic and NAI across a range of inter-related imaging modalities and category-selective regions in both the left and right hemispheres. Elucidating more integrated, individual-level neural associations of core social functioning in autism will pave the way for further work on identifying more fine-grained stratification, mechanistic and prognostic biomarkers, and the development of more personalised support. ### Competing Interest Statement JKB has been a consultant to, advisory board member of, and a speaker for Takeda/Shire, Medice, Roche, and Servier. He is not an employee of any of these companies and not a stock shareholder of any of these companies. He has no other financial or material support, including expert testimony, patents, or royalties. CFB is director and shareholder in SBGneuro Ltd. TC has received consultancy from Roche and Servier and received book royalties from Guildford Press and Sage. TB served in an advisory or consultancy role for ADHS digital, Infectopharm, Lundbeck, Medice, Neurim Pharmaceuticals, Oberberg GmbH, Roche, and Takeda. He received conference support or speaker fee by Medice and Takeda. He received royalities from Hogrefe, Kohlhammer, CIP Medien, Oxford University Press; the present work is unrelated to these relationships. The other authors report no biomedical financial interests or potential conflicts of interest. ### Funding Statement This project has received funding from the Innovative Medicines Initiative 2 Joint Undertaking under grant agreement No 115300 (for EU-AIMS) and No 777394 (for AIMS-2-TRIALS). This Joint Undertaking receives support from the European Unions Horizon 2020 research and innovation programme and EFPIA and AUTISM SPEAKS, Autistica, SFARI. Any views expressed are those of the author(s) and not necessarily those of the funders. This work was also supported by the Netherlands Organization for Scientific Research through Vidi grants (Grant No. 864.12.003 [to CFB]; from the FP7 (Grant Nos. 602805) (AGGRESSOTYPE) (to JKB), 603016 (MATRICS), and 278948 (TACTICS); and from the European Communitys Horizon 2020 Programme (H2020/2014-2020) (Grant Nos. 643051 [MiND] and 642996 (BRAINVIEW). This work received funding from the Wellcome Trust UK Strategic Award (Award No. 098369/Z/12/Z) and from the National Institute for Health Research Maudsley Biomedical Research Centre (to DM). DLF is supported by funding from the European Unions Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 101025785. EJHJ and RH received funding from SFARI GAIINS (grant number 10039678). SB-C is funded by the Autism Research Trust, the Wellcome Trust, the Templeton World Charitable Foundation and by the NIHR Biomedical Research Centre in Cambridge, during the period of this work. BHV is supported by the Swiss National Science Foundation [10001C_197480]. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: At each site, an independent ethics committee approved the study; all participants (where appropriate) and their parent/legal guardian provided written informed consent. The ethics committee (London Queen Square Health Research Authority Research Ethics Committee) of KCL and University of Cambridge gave ethical approval for this work (13/LO/1156). The ethics committee (Radboud Universitair Medisch Centrum Instituut Waarborging Kwaliteit en Veiligheid Commissie Mensgebonden Onderzoek and Regio Arnhem-Nijmegen (Radboud University Medical Centre Institute Ensuring Quality and Safety Committee on Research Involving Human Subjects Arnhem-Nijmegen) of Radboud UMC gave ethical approval of this work (2013/455). The ethics committee of Central Institute of Mannheim (UMM Universitaetsmedizin Mannheim, Medizinische Ethik Commission II (UMM University Medical Mannheim, Medical Ethics Commission II) gave ethical approval for this work (2014-540N-MA). The ethics committee of University of Rome (Universita Campus Bio-Medica De Roma Comitato Etico (University Campus Bio-Medical Ethics Committee De Roma) gave ethical approval of this work (8/14 PAR ComET CBM). I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Data produced in the present study are available upon reasonable request to the authors. Code is available at: https://github.com/allera/Llera\_elife\_2019\_1/tree/master/matlab\_flica_toolbox and https://github.com/amarquand/PCNtoolkit.