202 Chapter 7 analyses to assess the overall time-on-task effect per group and to compare both groups. The within blocks time-on-task effect was studied combining the two difficulty levels by creating two first-level contrasts: early half block > late half block; late half block > early half block. These contrasts were analysed per group and were compared between groups in second-level analyses. Post-hoc analyses were conducted by separating the moderate and higher difficulty levels as significant differences were seen between groups within blocks with the two difficulty levels combined. Error-related effect Block design analyses were implemented to compare error-related brain activity between people with narcolepsy type 1 and healthy controls as an event-related approach was not justified due to insufficient occurrence of errors. Error rates were calculated per participant for the two difficulty levels combined. Scores were added as covariates of interest in the contrast: task > baseline. Resulting contrasts were compared between the groups. Statistical analyses Statistical processing of demographic and task performance measures was done in IBM SPSS Statistics version 24 (International Business Machines Corporation, Armonk, NY, USA), with p-values < 0.05 considered statistically significant. Student’s t-tests were used for age, IQ, ESS, and behavioural SART performance comparisons between groups and a chi-squared test to identify possible sex differences. Overall SART, time-on-task over repetitions, and time-on-task within blocks performance measures were analysed using repeated measures ANOVAs. Additional paired-samples t-tests were conducted on overall SART performance to analyse within-participant effects per group. Student’s t-tests were used to study differences per difficulty level. Bonferroni correction for multiple comparisons thresholds were calculated using Simple Interactive Statistical Analysis (SISA) separately for error scores (correlation = 0.65, N = 9, p < 0.024) and reaction times (correlation = 0.86, N = 9, p < 0.037). The main task, time-on-task and error-related fMRI effects were determined per group using one-sample t-tests. For group comparisons, two-sample t-tests were used. The main task effect analyses per group were controlled for type I errors using family-wise error (FWE) correction. Cluster-correction for multiple comparisons was used for main task effect group comparisons and all time-on-task and error-related fMRI analyses. A p < 0.05 threshold and
RkJQdWJsaXNoZXIy MjY0ODMw