A Novel Machine Learning Algorithm to Predict the Lewy Body Dementias

A Novel Machine Learning Algorithm to Predict the Lewy Body Dementias Using Clinical and Neuropsychological Scores

Parkinson's disease dementia (PDD) and Dementia with lewy bodies (DLB) are dementia syndromes that overlap in many clinical features, making their diagnosis difficult in clinical practice, particularly in advanced stages. We propose a machine learning algorithm, based only on non-invasively and easily in-the-clinic collectable predictors, to identify these disorders with a high prognostic performance.

The algorithm will be develop using dataset from two specialized memory centers, employing a sample of PDD and DLB subjects whose diagnostic follow-up is available for at least 3 years after the baseline assessment. A restricted set of information regarding clinico- demographic characteristics, 6 neuropsychological tests (mini mental, PD Cognitive Rating Scale, Brief Visuospatial Memory test, Symbol digit written, Wechsler adult intelligence scale, trail making A and B) was used as predictors. Two classification algorithms, logistic regression and K-Nearest Neighbors (K-NNs), will be investigated for their ability to predict successfully whether patients suffered from PDD or DLB.

Inclusion Criteria:

the PDD group comprised of patients fulfilling the Criteria for probable PDD of the Movement Disorders Society (b) the DLB group comprised of patients, according to the recent revised criteria for probable DLB .

Exclusion Criteria:

• major psychiatrics disorders, depression