AUTHOR=Lock Christine , Tan Nicole Si Min , Long Ian James , Keong Nicole C. 

TITLE=Neuroimaging data repositories and AI-driven healthcare—Global aspirations vs. ethical considerations in machine learning models of neurological disease

JOURNAL=Frontiers in Artificial Intelligence

VOLUME=Volume 6 - 2023

YEAR=2024

URL=https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2023.1286266

DOI=10.3389/frai.2023.1286266

ISSN=2624-8212

ABSTRACT=Neuroimaging data repositories are data-rich resources that comprise brain imaging with associated clinical and biomarker data, hosted for the purposes of collaboration in the pursuit of Open Science. The potential for such repositories to transform healthcare is tremendous, both in their egalitarian approach of accessibility to researchers globally, and in their capacity to support machine learning (ML) and artificial intelligence (AI) tools. However, as ML/AI innovations have recently demonstrated, via generative AI and large language model applications, such tools have the potential to cause harms. These harms reflect wider challenges within societies to address bias and discrimination against groups who may be made more vulnerable, should bias be incorporated into methodologies utilized in the delivery of healthcare. Whilst these initiatives provide equitable access for scientists from high-, middle-and low-income countries, their knowledge harvest will be most applicable to their sponsors, reflecting the composition of their patient cohorts. Current discussions about the generalizability of ML models in healthcare provoke concerns of risk of bias -ML models are known to underperform in women and ethnic and racial minorities. ML approaches may therefore exacerbate existing healthcare disparities or cause post-deployment harm, i.e. by a biased prediction model deployed into clinical settings. Do neuroimaging data repositories and their capacity to support ML/AI-driven clinical discoveries, have both the potential to accelerate innovative medicine and harden the gaps of social inequities in the provision of neuroscience-related healthcare? In this paper, we examined the ethical concerns of ML-driven modelling of global community neuroscience needs arising from the use of data amassed within neuroimaging data repositories. We explored this in two parts; firstly, we conducted a theoretical experiment to argue for a South East Asian-based repository to redress global imbalances. We considered the ethical framework towards the inclusion vs. exclusion of the migrant worker population, a group subject to healthcare inequities. We then performed a mini hypothetical experiment; by using ML methodologies, we examined the impact of proportionality of patient cohorts within a pilot COVID-19 dataset towards altering structural brain findings. Finally, we proposed that such an approach may contribute towards a new strategy of promoting AI ethics.