Our technology is built on rigorous scientific research. Every claim is backed by peer-reviewed publications in leading medical informatics journals.
87%
Depression Detection
76%
Anxiety Detection
4
Published Studies
83.1%
CNN Model Accuracy
Peer-Reviewed
JMIR Formative Research, 2022
A Machine Learning Approach for Detecting Digital Behavioral Patterns of Depression Using Nonintrusive Smartphone Data
Choudhary S, Thomas N, Ellenberger J, Srinivasan G, Cohen R
This study demonstrates a novel mental behavioral profiling metric (the Mental Health Similarity Score), derived from analyzing passively monitored, private, and nonintrusive smartphone use data, to identify and track depressive behavior and its progression.
Key Findings
87% accuracy in detecting depression severity (PHQ-9 binary model)
558 participants in observational study over avg. 10.7 days
37 digital behavioral markers extracted from passive smartphone data
Significant Pearson correlation (r=0.73) with PHQ-9 questions 2, 6, and 9
Peer-Reviewed
JMIR Medical Informatics, 2022
A Machine Learning Approach for Continuous Mining of Nonidentifiable Smartphone Data to Create a Novel Digital Biomarker Detecting Generalized Anxiety Disorder
Choudhary S, Thomas N, Alshamrani S, Srinivasan G, Ellenberger J, Nawaz U, Cohen R
This study evaluates the accuracy of a novel mental behavioral profiling metric derived from smartphone usage for the identification and tracking of generalized anxiety disorder (GAD).
Key Findings
76% accuracy in tracking daily anxiety levels (GAD-7 binary model)
229 participants over an average of 14 days
34 digital behavioral features mined from continuous smartphone data
AUC of 80% for the binary XGBoost classification model
Preprint
JMIR (Preprint), 2024
Deep CNN-Based Continuous Monitoring of Depression Using Smartphone Usage Patterns: Algorithm Development and Validation
Srinivasan G, Trzesiok A, Chowdhary S, Cohen R, Ellenberger J
This study introduces a convolutional neural network (CNN)-based framework to predict continuous PHQ-9 scores, enable symptom-specific analysis, and provide confident classifications of depression severity using passive smartphone data.
Key Findings
83.1% overall accuracy with CNN-based framework
90.3% precision for positive depression cases
Continuous PHQ-9 score prediction (not just binary)
95% of cases confidently categorized with only 5% in uncertain range
Peer-Reviewed
IntechOpen - Digital Mental Health & Technology, 2022
The Importance of Using Binary Classification Models in Predicting Depression from a Machine Learning Perspective
Choudhary S, Srinivasan G
This article discusses the advantages of binary classification models in digital phenotyping for mental health, demonstrating that binary differentiation between presence versus absence of a condition is more reliable than multi-class severity determination.
Key Findings
Analysis of binary vs. multi-class classification for digital phenotyping
Binary models achieve significantly higher accuracy than multi-class
Addresses bias-creep from questionnaire-based ground truth
Demonstrates 87% depression and 76% anxiety detection using non-private data
We use cookies to enhance your browsing experience and analyze site traffic. By clicking "Accept", you consent to our use of cookies in accordance with our Privacy Policy and Terms of Service.
