Differentiation between atypical anorexia nervosa and anorexia nervosa using machine learning.

OBJECTIVE: Body mass index (BMI) is the primary criterion differentiating anorexia nervosa (AN) and atypical anorexia nervosa despite prior literature indicating few differences between disorders. Machine learning (ML) classification provides us an efficient means of accurately distinguishing between two meaningful classes given any number of features. The aim of the present study was to determine if ML algorithms can accurately distinguish AN and atypical AN given an ensemble of features excluding BMI, and if not, if the inclusion of BMI enables ML to accurately classify between the two. METHODS: Using an aggregate sample from seven studies consisting of individuals with AN and atypical AN who completed baseline questionnaires (N = 448), we used logistic regression, decision tree, and random forest ML classification models each trained on two datasets, one containing demographic, eating disorder, and comorbid features without BMI, and one retaining all features and BMI. RESULTS: Model performance for all algorithms trained with BMI as a feature was deemed acceptable (mean accuracy = 74.98%, mean area under the receiving operating characteristics curve [AUC] = 74.75%), whereas model performance diminished without BMI (mean accuracy = 59.37%, mean AUC = 59.98%). DISCUSSION: Model performance was acceptable, but not strong, if BMI was included as a feature; no other features meaningfully improved classification. When BMI was excluded, ML algorithms performed poorly at classifying cases of AN and atypical AN when considering other demographic and clinical characteristics. Results suggest a reconceptualization of atypical AN should be considered. PUBLIC SIGNIFICANCE: There is a growing debate about the differences between anorexia nervosa and atypical anorexia nervosa as their diagnostic differentiation relies on BMI despite being similar otherwise. We aimed to see if machine learning could distinguish between the two disorders and found accurate classification only if BMI was used as a feature. This finding calls into question the need to differentiate between the two disorders.

Artifact Adaptive Ideal Filtering of EMG Signals Contaminated by Spinal Cord Transcutaneous Stimulation.

The aims of this study are to characterize the contamination of EMG signals by artifacts generated by the delivery of spinal cord transcutaneous stimulation (scTS) and to evaluate the performance of an Artifact Adaptive Ideal Filtering (AA-IF) technique to remove scTS artifacts from EMG signals. METHODS: In five participants with spinal cord injury (SCI), scTS was delivered at different combinations of intensity (from 20 to 55 mA) and frequencies (from 30 to 60 Hz) while Biceps Brachii (BB) and Triceps Brachii (TB) muscles were at rest or voluntarily activated. Using a Fast Fourier Transform (FFT), we characterized peak amplitude of scTS artifacts and boundaries of contaminated frequency bands in the EMG signals recorded from BB and TB muscles. Then, we applied the AA-IF technique and the empirical mode decomposition Butterworth filtering method (EMD-BF) to identify and remove scTS artifacts. Finally, we compared the content of the FFT that was preserved and the root mean square of the EMG signals (EMGrms) following application of the AA-IF and EMD-BF techniques. RESULTS: Frequency bands of ~2Hz width were contaminated by scTS artifact at frequencies nearby the main frequency set for the stimulator and its harmonics. The width of the frequency bands contaminated by scTS artifacts increased with current intensity delivered using scTS ( [Formula: see text]), was lower when EMG signals were recorded during voluntary contractions compared to rest ( [Formula: see text]), and was larger in BB muscle compared to TB muscle ( [Formula: see text]). A larger portion of the FFT was preserved using the AA-IF technique compared to the EMD-BF technique (96±5% vs. 75±6%, [Formula: see text]). CONCLUSION: The AA-IF technique allows for a precise identification of the frequency bands contaminated by scTS artifacts and ultimately preserves a larger amount of uncontaminated content from the EMG signals.