Open Access Open Access  Restricted Access Subscription Access

Clustering of Multi-Modal Dataset using Ensemble Densenets for Early Mild Cognitive Impairment


Affiliations
1 Department of Artificial Intelligence and Data Science, Dr. N.G.P. Institute of Technology, India

   Subscribe/Renew Journal


Early Mild Cognitive Impairment (EMCI) is a transitional phase between normal cognition (NC) and Alzheimer’s disease. Accurate detection of EMCI can be challenging due to its subtle manifestations. Traditional methods often struggle to differentiate EMCI from NC using neuropsychological tests alone, necessitating advanced techniques for effective classification. We employed Ensemble DenseNets to cluster a multi-modal dataset comprising neuropsychological tests and clinical data. Generalized Estimating Equations (GEE) were used to analyze changes over time across various cognitive tests. Our model demonstrated significant findings: MMSE showed a time effect (β = 0.151, p = 0.01) with a notable decline in EMCI compared to NC (β = -0.299, p = 0.001). STM also showed significant results (time β = 0.105, p < 0.001). In the CVVLT total recall test, a time effect (β = 1.263, p < 0.001) and a decline in EMCI (β = - 0.510, p = 0.003) were observed. The method effectively clustered EMCI with a high degree of accuracy, showcasing the robustness of Ensemble DenseNets for early detection.

Keywords

Early Mild Cognitive Impairment, Ensemble DenseNets, Neuropsychological Tests, Generalized Estimating Equations, Cognitive Clustering
Subscription Login to verify subscription
User
Notifications
Font Size

Abstract Views: 90




  • Clustering of Multi-Modal Dataset using Ensemble Densenets for Early Mild Cognitive Impairment

Abstract Views: 90  | 

Authors

A.S. Shanthi
Department of Artificial Intelligence and Data Science, Dr. N.G.P. Institute of Technology, India

Abstract


Early Mild Cognitive Impairment (EMCI) is a transitional phase between normal cognition (NC) and Alzheimer’s disease. Accurate detection of EMCI can be challenging due to its subtle manifestations. Traditional methods often struggle to differentiate EMCI from NC using neuropsychological tests alone, necessitating advanced techniques for effective classification. We employed Ensemble DenseNets to cluster a multi-modal dataset comprising neuropsychological tests and clinical data. Generalized Estimating Equations (GEE) were used to analyze changes over time across various cognitive tests. Our model demonstrated significant findings: MMSE showed a time effect (β = 0.151, p = 0.01) with a notable decline in EMCI compared to NC (β = -0.299, p = 0.001). STM also showed significant results (time β = 0.105, p < 0.001). In the CVVLT total recall test, a time effect (β = 1.263, p < 0.001) and a decline in EMCI (β = - 0.510, p = 0.003) were observed. The method effectively clustered EMCI with a high degree of accuracy, showcasing the robustness of Ensemble DenseNets for early detection.

Keywords


Early Mild Cognitive Impairment, Ensemble DenseNets, Neuropsychological Tests, Generalized Estimating Equations, Cognitive Clustering