SPEECH impairments are often the earliest indicators of Parkinson’s disease , a rapidly growing neurological disorder affecting over 8.5 million people worldwide. Recent research by Middle Technical University (MTU) in Baghdad and the University of South Australia (UniSA) highlights the potential of artificial intelligence (AI) to revolutionise the diagnosis and monitoring of Parkinson’s disease through voice analysis. These findings were presented at the Fifth Scientific Conference for Electrical Engineering Techniques Research (EETR2024).
To explore AI’s capabilities, researchers reviewed advanced machine learning and deep learning techniques that analyse voice recordings for early signs of Parkinson’s. These methods use large datasets of audio samples from individuals with and without Parkinson’s to train AI algorithms to detect disease-specific vocal changes. Key features such as pitch, articulation, and speech rhythm are extracted and processed to categorise voice recordings. In one study, models demonstrated an impressive accuracy rate of 99%, underscoring their diagnostic potential. The study also addressed the benefits of AI for remote patient monitoring, reducing the need for in-person evaluations while offering consistent data to track disease progression.
The findings emphasise the transformative potential of AI-powered voice analysis in clinical practice. Early diagnosis could enable timely interventions, improving patients’ quality of life and slowing symptom progression. Despite these promising results, researchers stress the importance of further studies involving larger and more diverse populations to enhance AI model reliability. Additionally, integrating these tools into healthcare systems will require robust validation and training for clinicians.
Looking ahead, AI’s role in diagnosing and monitoring Parkinson’s disease is poised to grow, complementing neurologists’ efforts in early detection and personalised care. Expanding this technology’s reach could significantly impact how neurodegenerative diseases are managed, promoting efficiency and accessibility in healthcare systems worldwide.
Reference
Mujtaba H A et al. Parkinson’s disease detection from voice using artificial intelligence techniques: a review. AIP Conf. Proc. 2024;3232:040010.
