A New App for At-Home Cognitive Training: Description and Pilot Testing on Patients with Multiple Sclerosis

Abstract

Background: Cognitive impairment is common in people with neurological diseases and severely affects their social and professional life. It has been shown that intensive and personalized cognitive rehabilitation (CR), based on working memory exercises, leads to improved cognitive status of healthy and cognitive-impaired subjects. New technologies would help to promote accessible, at-home, and self-managed CR interventions.

Objective: The aim of this paper is to describe the design of Cognitive Training Kit (COGNI-TRAcK), an app for mobile devices, to self-administer an at-home, intensive, and personalized CR intervention based on working memory exercises, and test its disposability-to-use (usability, motivation to use, compliance to treatment) on cognitive-impaired patients with multiple sclerosis (MS).

Methods: COGNI-TRAcK includes user-friendly interfaces for personal data input and management and for CR intervention configurations. Inner routines automatically implement adaptive working load algorithms and allow data processing and analysis. A dedicated team developed COGNI-TRAcK with C# programming language, by using the platform Xamarin Studio 4.0.10 for Android (API level 15 and following). Three exercises based on working memory are now available. To assess the disposability-to-use of the system, patients with MS were selected as likely users due to the young age of disease onset. Cognitive-impaired patients with MS (N=16) with a mean age of 49.06 years (SD 9.10) and a mean score of 3.75 (SD 1.92) on the Expanded Disability Status Scale (EDSS) were submitted to an 8-week at-home intervention administered by the app. The intervention consisted of 5 daily scheduled 30-minute sessions per week. Disposability-to-use of COGNI-TRAcK was investigated by means of a questionnaire administered to patients at the end of the training.

Results: The adherence to the treatment was 84% (33.4/40). Of the patients with MS, 94% (15/16) understood the instructions given, 100% (16/16) felt independent to use COGNI-TRAcK at home, 75% (12/16) found the exercises interesting, and 81% (13/16) found the exercises useful and were motivated to use the app again. Moreover, during the exercises, patients with MS were highly motivated to perform well (mean score 3.19/4, SE 0.16), experienced rather low levels of stress (mean score 2.19/4, SE 0.26), were not bored (mean score 1.81/4, SE 0.30), and felt amusement (mean score 2.25/4, SE 0.23).

Conclusions: As COGNI-TRAcK is highly usable, motivating, and well-accepted by patients with MS, its effectiveness can now be investigated. To improve COGNI-TRAcK, new releases should contain more working memory exercises, have enhanced perceived amusement, and promote Internet communication procedures for data transfer and fostering remote control of the intervention.

Keywords: adaptive working load algorithms; cognitive impairment; cognitive rehabilitation; mobile device; mobile phone; self-management; tablet; usability; working memory.

Figure 1

An example of visuospatial working memory (Vs-WM) exercises in which patients have to remember a random sequence of 5 circles consecutively presented (the temporal order is defined following the direction of the arrows). Task-specific parameters for Vs-WM include grid size (minimum 2×2; maximum not defined), number of stimuli composing the sequence (minimum 1; maximum not defined), and the rate of stimuli presentation. Adaptive working load algorithms could operate on all the task-specific parameters to increment the difficulty level.

Figure 2

Examples of the operation N-back (Op-NB) exercises. The upper panel represents the operation 0-back (N=0) level of difficulty, whereas operation 1-back (N=1) is shown in the bottom panel. All the digits from 1 to 9 can appear on the screen and, consequently the sums range from 2 to 18. However, it is possible to reduce the difficulty of the task by limiting the digit span. Besides digit span, other task-specific parameters have to be taken into account such as the value of N (minimum 0; maximum not defined) and the rate of stimulus presentation. In Op-NB, adaptive working load algorithms operate only on N and rate of stimulus presentation.

Figure 3

Examples of the dual N-back (D-NB) exercise. The upper panel shows the D-NB difficulty level when N=0 ("Dual" 0-back), whereas the lower panel shows N=1 ("Dual" 1-back). Here, patients have to watch the stimulus, memorize it, and touch the buttons corresponding to the digit and cell deferred by one new stimulus. Task-specific parameters are N (minimum 0; maximum not defined), and the rate of stimuli presentation, both susceptible to modifications by the adaptive working load algorithms.

Figure 4

Results from the questionnaire on the usability of COGNI-TRAcK. The columns show the percentages obtained in the first 5 questions referring to usability and motivation.

Figure 5

Results from the questionnaire on the motivation and compliance to use COGNI-TRAcK. The columns indicate the mean value obtained for compliance on a scale from 0 to 4.