Haoran Xie, PhD, EdD | Department of Computing and Decision Sciences, Lingnan University; Kee Lee Chou, PhD; Xiaohui Tao, PhD; Soman Elangovan, MBBS; Raj Gururajan, PhD
Competition Sponsor: Research Grants Council of the Hong Kong Special Administrative Region, China
Award Year: 2023
According to the World Health Organization, depression is a highly prevalent mental illness that causes serious harm on a global scale, affecting an estimated 3.8% of the population. The success rates of depression treatments are often unsatisfactory. Repetitive transcranial magnetic stimulation (rTMS) is known to be an effective clinical treatment for clinical depression. However, the operational parameters of rTMS treatment are chosen by clinical professionals largely on the basis of experience, leading to ineffective treatment outcomes and medical resource management. To overcome this limitation, it is vital to understand what types of patients are most likely to respond well to clinical rTMS treatment at the patient-specific level. In this proposed project, we will leverage the availability of Hong Kong and Australian rTMS intervention data, patient characteristic data, and the latest progress in artificial intelligence (AI) to develop an AI-empowered rTMS depression treatment recommendation model. The model will be based on data-driven progressive patient profiling, comprising three main steps: data collection and preprocessing, representation learning for patient profiles, and recommendation of personalized treatment. Specifically, self-supervised AI algorithms will transform the original patient data space into a latent representation space, in which similar patients can be accurately grouped, thereby facilitating personalized treatment planning. Through a randomized clinical trial, we expect to reliably deliver personalized treatment (e.g., number of pulses, amplitude, treatment frequency, and administration site) to improve the effectiveness of rTMS treatment.