Principal Component Analysis of the Modified Clinical Test Ofsensory Interaction in Healthy Adult Humans

A number of mechanisms and sensory systems in humans are associated with the maintenance of balance. Diagnosis and monitoring of balance dysfunctions could be assisted by exploring deviations of data recorded from patients with comparative or reference data from healthy individuals. To this effect, principal component analysis (PCA) was applied to accelerometry obtained time domain balance data. The data were recorded from 21 healthy adults (10 males and 11 females, mean age 24.5 years, standard deviation 4.0 years, mean height 173.6 cm, standard deviation 6.8 cm, and mean weight 72.7 kg, standard deviation 9.9 kg) in the medio-lateral (ML) and anterior- posterior (AP) directions. The subjects performed tasks specified in the modified clinical test of sensory interaction on balance (mCTSIB) while an accelerometry device was attached at their lower back, in the position of the iliac crest. Eighteen-time domain measures that quantified body's displacement, velocity and acceleration were obtained and processed using PCA. Based on the observations from PCA, further investigations were carried out on the root mean square (RMS) velocity using the Bland and Altman plots and other statistical related analysis. It was observed that the anterior and posterior directions were more sensitive to the absence or presence of balance sensory (visual, somatosensory and vestibular) inputs as compared to the mediolateral (ML) direction. A greater coherence in sway information was observed in the somatosensory system as compared to the visual and vestibular systems. There was more stability in the interaction between the somatosensory and the vestibular systems as compared to that of the visual and vestibular systems. The results obtained could be helpful to clinicians in balance related analysis and diagnosis.