Researchers at the Indian Institute of Science (IISc) are currently undertaking a project which aims to collect data for commonly observed neuro-degenerative disorders and develop models for diagnosis using state-of-the-art machine learning approaches.
The Learning and Extraction of Acoustic Patterns (LEAP) lab at the Department of Electrical Engineering has embarked on this project along with Aster CMI hospitals to collect data for neurodegenerative diseases. It envisages using the massive outreach of smartphone technologies among the population as part of the project.
High burden of non-communicable diseases
“The burden of non-communicable neurological disorders has more than doubled in India in the last 30 years mainly due to the ageing of the population. Currently, this has a 10% contribution to the total deaths in the country. Further, medical evidence indicates a substantial increase in this figure due to the rapid demographic and epidemiological transition,” Sriram Ganapathy Associate Professor, Electrical Engineering, LEAP lab told The Hindu.
According to Mr. Sriram, patients with neurodegenerative diseases like Parkinson’s, Dementia and Amyotrophic Lateral Ssclerosis (ALS), especially those from the rural communities, show up for analysis or scans only about 18 months after the onset of all these diseases.
One of the key factors that can reduce the impact is the early diagnosis of neurological disorders.
“In the direction of diagnosis, the first major step is the clinical analysis that may also include MRI scans to identify the neurological condition of the subject. There are two major concerns with the current convention; the first is the cost and the second is lack of wide-spread access to clinical expertise as well as MRI scanning devices. These disadvantages cause an inhibition in testing which leads to delayed diagnosis,” Mr Sriram said.
He said that neurodegenerative disorders like Parkinson’s disease are quite evident in various modalities, like facial expressions, speech production, gait patterns and finger movements, which are easily measurable using smartphones.
Layer of screening
“Our work aims to bring a layer of screening before the current convention of diagnosis, by exploiting the massive outreach of smartphone technologies among the population. The current smartphones are equipped with good quality microphones and camera devices that allow inexpensive and remote sensing of patient data. The ultimate aim is to offer the tool in a highly cost-effective, easy to use, and edge computing enabled platform for wide-spread deployment on mobile phones,” Mr Sriram said.
IISc has been provided with medical expertise by a team led by Dr. B. Lokesh, lead consultant Neurologist, Aster CMI. So far, they have sampled about 200 subjects for Parkinson’s disease and are in the process of analysing the data. This data has been primarily collected from patients in and around the city of Bengaluru.
Key goals of the project
Highlights
- To develop a data repository of rich audio-visual data pertaining to specific neuro-degenerative diseases like Parkinson’s disease.
- To design machine learning models with multi-modal data that can enable the screening of the disease using the data recorded through smartphones.
- Provide recommendation and monitoring tools for patients and for medical practitioners for subsequent tests and diagnosis.
