Prediction of Antipsychotic Drug Doses for BPSD in Alzheimer's Disease Using Deep Learning Techniques.
Authors
Affiliations (9)
Affiliations (9)
- Shanghai Mental Health Center, Shanghai Jiao Tong University School of Medicine, Shanghai 200030, China.
- Alzheimer's Disease and Related Disorders Center, Shanghai Jiao Tong University, Shanghai 200030, China.
- Shanghai Institute of Traditional Chinese Medicine for Mental Health, Shanghai 201108, China.
- Shanghai Clinical Research Center for Mental Health, Shanghai 201108, China.
- Shanghai Key Laboratory of Mental Disorders Translational Research, Shanghai 201108, China.
- School of Biomedical Engineering, ShanghaiTech University, Shanghai 201210, China.
- Brain Health Institute at National Center for Mental Disorder, Shanghai 200030, China.
- Department of Psychiatry, The First Affiliated Hospital of Xinjiang Medical University, Urumqi 830054, China.
- Furong Laboratory, Xiangya Hospital, Central South University, Changsha 410078, China.
Abstract
<b>Background/Objectives</b>: Antipsychotic dosing for behavioral and psychological symptoms of dementia (BPSD) in Alzheimer's disease remains empirical and variable. This study develops a deep learning model to predict individualized antipsychotic doses from structural MRI. <b>Methods</b>: A transfer learning approach with a cascaded ResNet (Cas-ResNet) was used. The model was first pre-trained on a large healthy aging dataset (CBMFM, <i>n</i> = 646) for brain age prediction, then fine-tuned on a BPSD dataset (SMHC, <i>n</i> = 86) to predict the defined daily dose (DDD) of antipsychotics. Model interpretability was performed using Grad CAM to identify predictive brain regions. <b>Results</b>: The proposed model achieved a mean absolute error of 0.19 and a Pearson correlation of 0.66 between predicted and actual doses, outperforming baseline 3DCNN, VGG, and DenseNet. Key contributing regions included the left inferior temporal gyrus, right parahippocampal gyrus, right putamen, left middle temporal gyrus, and left caudate. <b>Conclusions</b>: This proof-of-concept study demonstrates that deep learning can predict personalized antipsychotic doses from structural MRI, offering an objective tool to standardize BPSD pharmacotherapy and reduce empirical prescribing. The identified brain regions provide neurobiological insights into treatment response.