Comparison between a Paper-Pencil Version and Computerized Version for the Realization of a Neuropsychological Test: The Example of the Trail Making Test

Article type: Research Article

Authors: Sacco, Guillaume^{a; b; *} | Ben-Sadoun, Grégory^a | Bourgeois, Jérémy^a | Fabre, Roxane^a | Manera, Valeria^a | Robert, Philippe^{a; b}

Affiliations: [a] Université Côte d’Azur, CoBTeK lab IA, France | [b] Université Côte d’Azur, CHU Pôle Réhabilitation Autonomie Vieillissement, CMRR, France

Correspondence: [*] Correspondence to: Guillaume Sacco, SSR post-AVC Hôpital Cimiez, Pavillon Mossa 4ème étage aile Sud, 4 av Reine Victoria, CS 91179, 06003 Nice, France. Tel.: +33 609733254; E-mail: [email protected].

Abstract: Background:Neuropsychological tests are particularly important for the clinical evaluation and Alzheimer’s disease (AD) diagnosis. However, the tests currently employed for neuropsychological assessment have been developed several decades ago, and thus they do not fully exploit the potential provided by modern digital tools. One of the tests most commonly employed to assess attention and executive functions is the Trail Making Test (TMT). Objective:The aim of this study was to evaluate whether the TMT developed and used for the serious exergame X-Torp (TMTX-Torp) can be used to evaluate cognitive functions such as mental flexibility. Methods:Adjusted multivariate mixed model was used to compare performances in the TMTX-Torp to performances in the standard variant (TMTs) in three populations. 21 participants with AD (78.6y±8.5 y), 27 with mild cognitive impairment (MCI) (76.8y±8.5 y), and 27 healthy (HEC) (71.8y±7.4 y) were included. Results:A difference was observed for the TMT A between AD and HEC and for the TMT B between AD and MCI and between AD and HEC. Whatever the variant of the TMT, we found a positive linear correlation between the time to complete the TMTX-Torp and the TMTs for HEC (TMT A: r = 0.75, p < 0.001; TMT B: r = 0.52, p = 0.008) and MCI participants (TMT A: r = 0.53, p = 0.005; TMT B: r = 0.48, p = 0.025) but not for AD participants. Conclusion:Although these versions of the TMT were not identical, the results showed that both versions were able to discriminate between HEC, MCI, and AD populations.

Keywords: Alzheimer’s disease, executive function, neurocognitive disorders, serious games

DOI: 10.3233/JAD-180396

Journal: Journal of Alzheimer's Disease, vol. 68, no. 4, pp. 1657-1666, 2019