[Online/At Home Completion] Measurement of Cognitive Performance through Smartphone Tasks - Semester 2 Update
Background
Mild traumatic brain injury (mTBI) is the most common form of traumatic brain injury (TBI). Despite how common mTBI is, we know little about the trajectory of cognitive recovery during the first three months post injury. The overall aim of this project is to be able to measure cognitive performance using smartphones in healthy individuals, in order to investigate its potential for future use in tracking cognitive recovery in patients with mTBI.
Research Questions / Hypotheses
This study aimed to investigate the relationship between performance on a smartphone version of the n-back paradigm, adminstered via CheckCog, and laboratory tasks measuring working memory, response inhibition, and fluid reasoning. The research questiosn were: 1) is there a significant correlation between measures of performance on the smartphone and computerised benchtop version of the n-back? 2) Is there a significant correlation between performance on the smartphone n-back task and the measures of fluid reasoning and response inhibition?
Participants
16 REP participants were involved with this study in semester 1, 2020. Participants were required to own a smartphone in order to be eligible for inclusion in the study, with CheckCog available for both iPhone and Android. Exclusion criteria included individuals who fulfilled criteria for mTBI as outlined by the World Health Organisation Collaborating Centre Task Force, possessed any upper limb injury that may affect ability to complete smartphone tasks, and/or who had a current psychiatric diagnosis that could affect performance on cognitive tasks. Additionally, participants who had completed the following REP studies were not able to participate: 'Tracking Cognitive Recovery in Patients with Mild Traumatic Brain Injury Using Smartphones' and 'Cognition and drug use behaviour in the general population'.
Methods
Participation in this study involved three components: (1) Completion of a short online survey, (2) Completion of computerized cognitive tasks delivered via a desktop computer in the lab and (3) Downloading and using a freely available smartphone application (CheckCog) designed for this study. The first two components of the study took place in the Redmond Barry Building, University of Melbourne. It took approximately 90 minutes to complete these two components. (1) Participants completed a short online questionnaire (approximately 20 minutes) at the initial assessment time point. These questionnaires helped us to understand mood and health factors. (2) Participants then completed computerized cognitive tasks in the Redmond Barry Building, University of Melbourne (the n-back, stop signal task, and an online brief version of Raven’s advanced progressive matrices). (3) Participants were then instructed to download and use a freely available smartphone application designed for this study, and to complete one session of the short cognitive tasks using their smartphone at home in the days after. These cognitive tasks were shorter versions of the tasks initially completed on campus (15 minutes).
Results
In relation to demographics, descriptive statistics will be run on the data from the Qualtrics questionnaire to describe the overall sample. Correlational analyses were run to investigate the relationship between performance on the smartphone version of the n-back on the and the laboratory n-back, and also between the smartphone n-back and the measures of fluid reasoing and response inhibition, respectively. A significant moderate correlation was found between overall accuracy on the CheckCog and laboratory versions of the n-back. Non-significant correlations were demonstrated between the CheckCog n-back and laboratory measures of response inhibition and fluid intelligence.
Implications
This study provides encouraging initial evidence for the convergent validity of the ambulatory version of the n-back in a healthy community sample. The significant moderate correlation between total accuracy on the CheckCog and laboratory versions of the task is promising given the underpowered sample. These preliminary findings provide scope for future research to further elucidate these relationships through investigation of broader performance metrics, inclusion of more robust measures, and use of aggregated ecological momentary assessment. These findings will be communicated in an MPsych thesis.