sMRI-based Applications
Parkinson’s disease (PD) is the second-most common neurodegenerative disorder in the world and is often characterized by brain tissue atrophy. Its heterogeneity in clinical symptoms is clearly visible by its phenotypic representation such as overlapping motor and non-motor dysfunctions. Structural MRI (sMRI) is a medical imaging technique used to visualize the internal structures of the human body, primarily the brain. It is a non-invasive method that does not use ionizing radiation. sMRI is widely employed in research and clinical settings, offering detailed anatomical information and high contrast between grey and white matter of brain. Few problems related to PD studied by us using sMRI are below
Subtyping
Subtyping of Parkinson's disease (PD) is the process of dividing PD into subgroups based on shared characteristics to help with clinical care and research. Although several attempts have been made in the past to address PD heterogeneity, a stable subtyping approach to categorize the patients, is the need of the hour. Recently, attempts are being made to find subtypes within Parkinson’s disease using structural magnetic resonance imaging (sMRI) and data driven algorithms. We analysed brain grey matter information to find distinct subtypes and correlated them with the clinical features. Our study showed the deciphered subtypes had differences in connectivity pattern. Graph-theory based network analysis was used to obtain connectivity metrics. Three subtypes were found with differences in frontal and temporal gyrus regions of brain. Inter-subtype differences in network metrics were also observed. Thus, successful subtyping will not only help in clinical analysis, but also be useful in precision treatment.
UBNIN and Modified UBNIN Algorithms
Individuals, whether healthy or disease, have a specific pattern of brain network due to interindividual differences. So, now the question is- Can a brain network be encoded? To address this, novel algorithms called Unique Brain Network Identification Number (UBNIN) and Modified UBNIN were proposed for encoding brain networks of an individual subject. Each subject’s brain volume was parcellated from structural MRI scans and individual adjacency matrix was constructed. UBNIN was highly weighted on the last node while another variant, called Modified-UBNIN (UBNIN-MT,MN) algorithm was highly weighted on the node with the highest network degree (i.e., connections). From initial results we observed that the numerical representation of these algorithms seems to be distinct for each individuals brain network. These algorithms may be implemented as neural signature of an individual’s unique brain connectivity, thereby useful for brainprinting applications.
Brain Age prediction
Unlike chronological age, which measures a person’s age in years since birth, the biological age of the brain is estimated using its morphological, functional, or molecular properties of brain. Brain age is a widely employed metric to assess and quantify an individual’s brain health. Brain ageing is linked to cognitive decline and an increased risk of neurodegenerative diseases. The difference between the estimated brain age and the chronological age is called the brain age gap. We are using deep learning methods to decipher brain age of an individual from grey and white matter regions of sMRI
| Google Scholar | Research Gate | Github Repository |