Amalgamating artificial intelligence & neuropsychiatry: Multimodal approach for precision diagnosis

Abstract

Human body is an intricate system of complex associations between numerous chemicals, neurotransmitters incessantly striving towards homeostasis. This process is so tightly regulated that these processes quickly reverse the alterations caused due to any insults suffered by the organism; when these repeatedly attack the organism’s system, they culminate as diseases. Brain being the hub of such complex interplay, is the most vulnerable site for such imbalances. As per 2021 World Health Organisation Report, over 3 billion of the global population live with a neuropsychiatric ailment. Owing to the extremely limited options for diagnosis in neuropsychiatry, it is tedious to arrive at a definitive diagnosis early. Normally, diagnosis of neuropsychiatric disorders like autism spectrum disorder, generalized anxiety disorders, obsessive compulsive disorder, schizophrenia etc. demands persistence of symptoms for over 6 months with no body fluid test to appropriately quantify the diagnosis, it demands the need for a novel approach for bridging this significant gap. With the ever-advancing technological era, the advent of artificial intelligence has revolutionized the medical paradigm. Integration of such multimodal neuroimaging databases, algorithms, techniques has made individualized diagnosis & treatment of neuropsychiatric disorders plausible. Considering the colossal data revolving around artificial intelligence neuroimaging, particularly address disease-specific challenges & rarely collate how multimodal frameworks amalgamate neurology & psychiatric disorders. In addition, no prior literature systematically compares biomarker-rich & scare diseases & emphasizes the challenges faced by artificial intelligence among these groups. This review bridges this research gap by presenting an integrated, comprehensible manuscript which links multimodal networks of supervised, unsupervised, reinforcement learning models across the whole of neuropsychiatric spectrum. The novelty of this manuscript lies in highlighting the transdiagnostic perspectives of multimodal approaches of neuroimaging which are artificial intelligence-driven. This aids in filling the gaps between psychiatric & neurological disorders thereby practically focussing on the blueprint towards precision diagnostic models- an entirely understudied aspect in previous literatures. The primary aim of this manuscript is to explore the fusion between the multimodal approach of artificial intelligence & neuropsychiatric diagnosis, their implications, and challenges faced in this integration. We meticulously analyze the applications of artificial intelligence in precision medicine in the field of neuropsychiatry to ensure timely diagnosis and prevent complications, to improve patients’cost of living, prognosis as well as future directions of this amalgamation.