CONF Purohit_ICASSP_2025/IDIAP Automatic Parkinson's disease detection from speech: Layer selection vs adaptation of foundation models Purohit, Tilak Ruvolo, Barbara Orozco-Arroyave, Juan Rafael Magimai-Doss, Mathew EXTERNAL https://publications.idiap.ch/attachments/papers/2025/Purohit_ICASSP_2025.pdf PUBLIC Proceedings of IEEE International Conference Acoustics, Speech, and Signal Processing (ICASSP) 2025 In this work, we investigate Speech Foundation Models (SFMs) for Parkinson’s Disease (PD) detection. We explore two main approaches: using SFMs as frozen feature extractors and fine-tuning/adapting them for PD detection tasks. We propose a cross-validation based layer selection methodology to identify the layer effective for PD detection. Additionally, we compare the performance of the layer selection scheme with full fine- tuning and parameter-efficient fine-tuning (PEFT) using Low- Rank Adaptation (LoRA). Our results show that layer selection and LoRA based fine-tuning can perform on par with full fine- tuning, providing a more parameter-cost efficient alternative. The highest accuracy was achieved by fine-tuning Whisper with LoRA.