A comparison of assessment of cognitive functioning on phones and tablets

Introduction: In addition to clinical trials conducted at research institutes and hospitals, in some instances it is vital to conduct cognitive assessments on location. For example, naturalistic studies conducted at specific venues, or patients tested at home. In those instances, mobile device that assesses cognitive function may be an outcome. Over the past decade, several test batteries for mobile phones and tablets have been developed. One of these is the PenscreenSix test battery. Aim: The purpose of this study was to examine whether this test battery yields similar results on a mobile phone and a tablet. It was further investigated how many training sessions are required on both devices to attain baseline performance levels. Methods: N = 39 healthy volunteers participated in a 1-day study to compare cognitive test performance on a mobile phone and tablet. Participants completed five training sessions and one final session per device. The PenscreenSix test battery took 18 minutes to complete and consisted of six 3-minute tests, measuring attention (Number Pairs Test, NP and Arrow Flankers Test, AF), psychomotor functioning (Arrow reaction time test, AR), working memory (Memory scanning test, MS), paired associate learning (Shape pairs, SP), and comprehension Serial sevens, S7). Outcome measures for each test were mean reaction time (RT) and the percentage of errors. The training sessions were separated by 5-10 minute breaks. Perceived mental effort was assessed after each test session, using the Mental Effort Scale. Statistical analyses were conducted using SPSS, version 24. Mean RT from each training session was compared with the final session using a paired samples T Test. The percentage of errors of each training session was compared to the final session using the nonparametric paired samples Wilcoxon Signed Ranks Test. Results: On the phone, when compared to the final session, significantly slower RTs (p <0.05) were found on session 1 of the NP, MS, and SP test, AR session 4, S7 session 1, 2 and 3. On the tablet, when compared to the final session, significantly slower RTs (p <0.05) were found on session 1 of the NP, MS, SP, and S7 test, and MS and S7 session 2 and 4. On the phone, significant differences in percentage of errors with the final session were found for MS session 4, AR session 1 and 5, and S7 session 1, 4 and 5. On the tablet, significant differences in percentage of errors with the final session were found for AR session1, 4, and 5, S7 session 1, SP session 3 and 4, and NP session 1, 2 and 5. All other comparisons with the final session did not reach statistical significance. No significant differences were observed between the devices on the final test session. No significant effect on mental effort were found. Conclusion: Both tablets and phone tests are suitable to be used in clinical trials. However, it is essential to train subjects on the tests. Compared to phones, more training sessions are required on tablets to attain baseline performance levels.