Touchscreen typing pattern analysis for remote detection of the depressive tendency

Abstract

Depressive disorder (DD) is a mental illness affecting more than 300 million people worldwide, whereas social stigma and subtle, variant symptoms impede diagnosis. Psychomotor retardation is a common component of DD with a negative impact on motor function, usually reflected on patients’ routine activities, including, nowadays, their interaction with mobile devices. Therefore, such interactions constitute an enticing source of information towards unsupervised screening for DD symptoms in daily life. In this vein, this paper proposes a machine learning-based method for discriminating between subjects with depressive tendency and healthy controls, as denoted by self-reported Patient Health Questionnaire-9 (PHQ-9) compound scores, based on typing patterns captured in-the-wild. The latter consisted of keystroke timing sequences and typing metadata, passively collected during natural typing on touchscreen smartphones by 11/14 subjects with/without depressive tendency. Statistical features were extracted and tested in univariate and multivariate classification pipelines to reach a decision on subjects’ status. The best-performing pipeline achieved an AUC = 0.89 (0.72–1.00; 95% Confidence Interval) and 0.82/0.86 sensitivity/specificity, with the outputted probabilities significantly correlating (>0.60) with the respective PHQ-9 scores. This work adds to the findings of previous research associating typing patterns with psycho-motor impairment and contributes to the development of an unobtrusive, high-frequency monitoring of depressive tendency in everyday living.

Screening of Parkinsonian subtle fine-motor impairment from touchscreen typing via deep learning

“…This is important as, the validated clinical AD diagnosis tests are not sensitive enough in detecting deviations from healthy-ageing trajectory, and suffer from intrinsic biases 43 . Moreover, progressive slowing of fine motor dexterity has been associated with cognitive deficits 44 , which also allow the investigation of the current method for depressive tendency screening 27 , 45 and mood alterations associated with mental health 46 . In conclusion, remote passive screening of subtle fine-motor symptoms via deep learning-based smartphone typing information reveals great potential in estimating subtle FMI and its severity via natural use of smartphone.…”

Keystroke Dynamics Patterns While Writing Positive and Negative Opinions

“…Biometric methods based on keystroke dynamics may be applied in several areas, such as user authentication [ 3 , 4 , 5 , 6 ], emotion recognition [ 1 , 7 , 8 , 9 , 10 , 11 , 12 ], monitoring mood disorders [ 13 , 14 ] and so forth. The proposed solutions are usually based on hand-crafted features extracted on the basis of keystroke timing or frequency characteristics, for example, dwell time, flight time, typing speed, frequency of using selected keys and so forth.…”

K-EmoCon, a multimodal sensor dataset for continuous emotion recognition in naturalistic conversations

“…Then why do machines need emotional skills? With advances in Machine Learning and Artificial Intelligence, the transition from human to machine is noticeable in all areas of the society, including those requiring expertise such as medical prognosis/diagnosis 5 , 6 or automobile driving 7 . It seems inevitable that these narrow AI systems 8 supersede human experts in respective domains, as it has already been demonstrated with AlphaGo’s superior performance in the game of Go over human champions 9 , 10 .…”