Abstract— Multiple Sclerosis (MS) is the most common cause, (after trauma) of neurological disability in young adults in Western countries. While several Magnetic Resonance Imaging (MRI) studies have demonstrated a strong association between the presence of cortical grey matter atrophy and the progression of neurological impairment in MS patients, the neurobiological substrates of cortical atrophy in MS, and in particular its relationship with white matter (WM) and cortical lesions, remain unknown. The aim of this study was to investigate the interplay between cortical atrophy and different types of lesions at Ultra-High Field (UHF) 7 T MRI, including cortical lesions and lesions with a susceptibility rim (a feature which histopathological studies have associated with impaired remyelination and progressive tissue destruction). We combined lesion characterization with a recent machine learning (ML) framework which includes explainability, and we were able to predict cortical atrophy in MS from a handful of lesion-related features extracted from 7 T MR imaging. This highlights not only the importance of UHF MRI for accurately evaluating intracortical and rim lesion load, but also the differential contributions that these types of lesions may bring to determine disease evolution and severity. Also, we found that a small subset of features [WM lesion volume (not considering rim lesions), patient age and WM lesion count (not considering rim lesions), intracortical lesion volume] carried most of the prediction power. Interestingly, an almost opposite pattern emerged when contrasting cortical with WM lesion load: WM lesion load is most important when it is small, whereas cortical lesion load behaves in the opposite way. Clinical Relevance— Our results suggest that disconnection and axonal degeneration due to WM lesions and local cortical demyelination are the main factors determining cortical thinning. These findings further elucidate the complexity of MS pathology across the whole brain and the need for both statistical and mechanistic approaches to understanding the etiopathogenesis of lesions. I. INTRODUCTION Multiple sclerosis (MS) is one of the most common causes of neurological disability in young adults in the Western world [1]. Different radiological features, such as brain Magnetic Resonance Imaging (MRI) “demyelinating” lesions and grey matter (GM) atrophy, are commonly used to diagnose and evaluate disease progression in MS patients [2]. * This work was supported by grants from the National Multiple Sclerosis Society (NMSS 4281-RG-A1 and NMSS RG 4729A2/1), National Institutes of Health R01NS078322-01-A1, and United States Army W81XWH-13-1- 0122. # These authors contributed equally to this work A.C. and N.T. are with the Department of Biomedicine and Prevention, University of Rome “Tor Vergata”, C.A.T., A. M., V.T.B. and C.M. are with Several MRI studies have shown that GM atrophy arises early in the course of the disease and accelerates with disease progression [3], and additional studies have also demonstrated a strong association between GM atrophy and neurological impairment in MS patients [4] evaluable through functional MRI [5]. As cortical atrophy seems to be the main driver of GM atrophy [6], the understanding of the neurobiological substrates and main determinants of cortical atrophy in MS could be instrumental in predicting disease progression and stratifying the disease subtypes. Cortical (both intracortical and leukocortical) demyelinated lesions constitute a substantial part of the total lesion load in MS brain [7]. In addition, MS patients may exhibit chronically active white matter lesions, which are identifiable on susceptibility-weighted MR images by their characteristic paramagnetic rim (commonly called “rim lesions”) [8], [9]. While both cortical and rim lesions load as well as cortical atrophy are relevant for the diagnosis and the evaluation of MS progression, very little is known about their interplay and, in particular, about how the differential occurrence of one or more types of lesion may be related to cortical atrophy. In this context, it is not clear whether cortical atrophy is mainly the result of local pathological processes or, instead, disconnection from other brain regions which may result by the disruption caused by white matter (WM) lesions. A strong limitation in the investigation of this question is the ability to actually detect and differentiate cortical and rim lesions (as well as of evaluating their extension) when employing MR scanners equipped with static fields with intensities typically found in clinical centers (3T or even 1.5T) [10]. This often allows clinicians to detect and evaluate only a small portion of lesions. In this context, recent studies have demonstrated that ultra-high field (UHF) human MRI (7T) significantly improves in vivo imaging of both cortical and rim lesions in MS patients [11], [12]. UHF radiological findings are therefore of strong clinical relevance and may represent the best candidates for investigating the differential role of all lesion types in the progression of cortical atrophy in patients affected by MS. The aim of this study was to understand the interplay between cortical thickness and different types of lesions, by leveraging radiological markers (cortical and rim lesion load Massachusetts General Hospital, Boston, United States. C.M., A. M., V.T.B and N.T. are also with A. A. Martinos Center for Biomedical Imaging, Boston, United States. Corresponding author: Allegra Conti. E-mail: allegra.conti@uniroma2.it. An interpretable machine learning model to explain the interplay between brain lesions and cortical atrophy in multiple sclerosis A. Conti, C.A. Treaba, A. Mehndiratta, V.T. Barletta, C. Mainero # , and N. Toschi # , Senior Member, IEEE 2021 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) Oct 31 - Nov 4, 2021. Virtual Conference 978-1-7281-1178-0/21/$31.00 ©2021 IEEE 3757