Comparing online and face-to-face administration of a neuropsychological computerized attention test: Assessment modality does not influence performance.

Background: The cognitive impairment associated with the COVID-19 pandemic highlighted the need for teleneuropsychology (1). Moreover, neurologic diseases associated with mental deterioration usually require the use of the same neuropsychological instrument to assess cognitive changes across time. Therefore, in such cases, a learning effect upon retesting is not desired. Attention and its subdomains can be measured using Go/no-go tests, such as, the Continuous Visual Attention Test (CVAT). Here, we administered the CVAT to investigate the effect of modality (online vs. face-to-face) on attentional performance. The variables of the CVAT measures four attention domains: focused-attention, behavioral-inhibition, intrinsic-alertness (reaction time, RT), and sustained-attention (intra-individual variability of RTs, VRT).

Methods: The CVAT was applied face-to face and online in 130 adult Americans and 50 adult Brazilians. Three different study designs were used: (1) Between-subjects design: healthy Americans were tested face-to-face ( = 88) or online ( = 42). We verified if there were any differences between the two modalities. (2) Within-subjects design: Brazilians participants ( = 50) were tested twice (online and face-to-face). For each CVAT variable, repeated measures ANCOVAs were performed to verify whether modality or first vs. second tests differ. Agreement was analyzed using Kappa, intraclass correlation coefficients, and Bland-Altman plots. (3) Paired comparisons: we compared Americans vs. Brazilians, pairing subjects by age, sex, and level of education, grouping by modality.

Results: Assessment modality did not influence performance using two independent samples (between-subjects design) or the same individual tested twice (within-subjects design). The second test and the first test did not differ. Data indicated significant agreements for the VRT variable. Based on paired samples, Americans did not differ from Brazilians and a significant agreement was found for the VRT variable.

Conclusion: The CVAT can be administered online or face-to-face without learning upon retesting. The data on agreement (online vs. face-to-face, test vs. retest, Americans vs. Brazilians) indicate that VRT is the most reliable variable.

Limitations: High educational level of the participants and absence of a perfect balanced within-subjects design.