Nadi Tarangini Pulse Patterns in type 2 Diabetes Mellitus

Abstract

Background: Ayurveda is defined as the “Science of Life”. Medicine is one of the important sub-parts of Ayurveda. Roga and Rogi parikṣa was given the utmost importance, and in it Naḍi parikṣa (pulse based diagnosis) is considered as the foremost examination method in aṣṭavidha rogi parikṣa for assessing the healthy state, diagnosis and prognosis of the disease. Aim and Objectives: The aim of the study is to differentiate the pulse waveforms in Non Diabetes, Pre-diabetes and Type 2 Diabetes Mellitus, individuals using Naḍi Tarangini Instrument. Objective is to study the pulse wave forms in diabetics, pre-diabetics and non diabetics and to compare the doṣa predominance in them according to āyurvedic concept of naḍi parikṣa. Methodolgy and Design: All the volunteers (individuals/ patients) from the Stop Diabetes Camp (SDM) in Rajkot, Udaipur and Chittorgarh 2013. (n=376) age ranging between 30-70yrs were screened using American Diabetes Association (ADA) diabetes risk test. Along with medical information their pulse waveforms were recorded using Nāḍī Tarangini and subsequently analyzed for pattern recognition. Results: Data analysis showed vāta predominant signals in 309 subjects. All the pulse signals are first provided as input to the feature extraction methods of Fourier transform, wavelet transform and auto-regressive modeling. The resulting features are used in the random forest classifier. The random forest classifier is implemented in Weka with parameters as 'unlimited' number of trees and depth of 10.Out of the whole dataset, approximated two third of the randomly chosen data was used as a training set and remaining one third of the data was used as a testing set. And the classification was performed for three sets Non Diabetes (ND), Pre Diabetes (PD) and Type 2 Diabetes Mellitus (DM). The 10-fold cross validation accuracy of the classification process is 86.84%. The precision and recall numbers were got during the classification, which showed high precision in T2DM with 95.24%, indicating that the classification process returned substantially more relevant results than irrelevant. It’s very important in the detection and diagnosis of diabetes. Conclusion: The high precision percentage in diabetes group revealed vāta doṣa predominace in Type 2 Diabetes Mellitus during the Nāḍī parikṣā using Nāḍī Tarangini instrument. Thus this instrument can be a reliable diagnostic tool. Further studies are warranted in this regard.