JMIR Publications

ABSTRACT

Background:

Background:

Typing on the smartphone has become a near daily activity requiring upper limb and cognitive function. Analysis of keystroke dynamics (KD) from regular typing can therefore potentially be utilized for passive and continuous monitoring of functioning in patients with multiple sclerosis (MS).

Objective:

To study the longitudinal association between KD and clinical arm function and cognition outcomes in MS.

Methods:

During a 1-year follow-up, 102 MS patients were assessed 3-monthly on arm function (Nine-Hole Peg Test, NHPT) and information processing speed (Symbol Digit Modalities Test, SDMT) and KD data was continuously obtained from regular typing on the participants’ own smartphone. Press and release latencies of alphanumeric keys constituted the Fine Motor Score Cluster (FMSC), latencies relating to punctuations and backspaces constituted the Cognition Score Cluster (CSC). The longitudinal association between clinical outcomes and keystroke clusters was assessed with linear mixed models. Between- (βbetween) and within- (βwithin) subjects effects were distinguished.

Results:

Mean±SD scores of FMSC and CSC were 0.43±0.16 and 0.94±0.41 seconds, respectively. FMSC significantly associated with NHPT: βbetween=15.9, 95% CI [12.2, 19.6] and βwithin=6.9, 95% CI [2.0, 11.9]. CSC significantly associated with SDMT between subjects (βbetween=-11.2, 95% CI [-17.3, -5.2]), but not within subjects (βwithin=-0.4, 95% CI [-5.6, 4.9]).

Conclusions:

Smartphone KD were longitudinally associated with MS outcomes. Worse arm function corresponded with longer latencies in typing both across and within patients. Worse processing speed corresponded with higher latencies relating to punctuations and backspaces across subjects. Hence, KD can be utilized as a potentially digital biomarker for remote monitoring and predicting clinical outcomes in MS. Clinical Trial: NL7070 (trialregister.nl)