Résumé
In this study, we participated in the validation of LAZ-50 for the diagnosis of fine speech disorders. This lexico-semantic battery containing six proofs with low-frequency words and idioms was designed and developed by Christelle Bolloré and Claire Sainson due to the lack of tools in this field. The objective is to participate in the validation of this battery on a population of subjects (n=18) with mild neurocognitive impairment following a head injury (n=9) or left-sided stroke (n=9), in order to ensure the good discrimination between patients and healthy subjects by analyzing the performances of patients and matched healthy volunteers on the different tests of the LAZ-50. Our results show first of all a good convergent validity of the battery. However, the operational hypotheses on discriminant validity suggesting that patients perform significantly worse than matched healthy volunteers could not be validated in patients with left-sided stroke (n=9): no significant difference in this direction was found. These same hypotheses are partially validated for brain injury patients (n=9) in the idiom tests: the idiom explanation score and the total score for idioms were significantly lower in patients. We also observe a high degree of internal consistency of the battery for the idiom part, and almost satisfactory for the word part. This is a preliminary study, and the need for a larger number of subjects to obtain a more representative sample of all patients concerned by these mild disorders argues in favor of continuing the study. The final objective is to make this battery available to objectify subtle lexical-semantic disorders in these populations, with the aim of proposing appropriate.