Resumo

VIVAS, Leticia Yanina. Application of a semantic network analysis method in stroke patients. Interdisciplinaria [online]. 2010, vol.27, n.1, pp.147-162. ISSN 1668-7027.

There are many ways in which semantic organization can be assessed. The methods usually used are useful in many ways but fail to provide the possibility to visualise the configuration of a semantic network corresponding to a certain group of concepts and to make qualitative and quantitative comparisons between different networks. The aim of the current research is to show the application of a semantic distance judgment method to stroke patients. This method is called DISTSEM and it was build on the basis of Spreading Activation Theory (Collins & Loftus, 1975). It allows obtaining the configuration of a semantic network according to the semantic distance judgment realized by the individuals between pairs of concepts of different and equal semantic category given by the examiner. The method has been administered to 30 stroke patients and 30 healthy controls. In the current paper, the response patterns of the patients are analyzed quantitatively and qualitatively and comparisons are made with the production of healthy controls. In the first place, all the matrixes were correlated with an interjudges matrix and the results show that strokepatients obtained significantly lower correlations than controls. In the second place, the matrixes of the stroke patients were correlated with a matrix that represented the mode of the estimations made by the group of healthy subjects. There were found two atypical groups. On the one hand, there was a group of patients that found scarce relations between concepts. On the other hand, there was another group that found a great number of relations between concepts, including relations between concepts that were not from the same semantic category. This data can be interpreted in light of Collins and Loftus theory as a failure of the spreading of the activation. Some patients seem to fail in the searching process within the network, particularly in the control of the spread of the activation necessary to look for the similarities requested in the task. In the first case, there was an inhibition in the activation spreading, which did not allow reaching the intersection node between the presented pair of concepts. In the second case, there was an activation spreading superior than expected for that task, that is to say that it failed the inhibition of the competitors that allow establishing the estimation accurate for the task (Howard, Nickel, Coltheart, & Cole-Virtue, 2006). Furthermore, a Multiple Correspondence Analysis was performed with the justifications used in each election. This analysis showed that there was a group of patients that tended to use atypical justifications in some pairs of concepts. Some of these patients justify their elections using functional justifications (e.g., "They are alike because they are useful for men"), another group used perceptual justifications (e.g., "They are alike because they have legs") and there was a small group that found scarce similarities. This indicates that the justifications used by this group of stroke patients tend to be less precise and less adequate for the context of the task. This inappropriate selection of responses leads to an explanation related to executive functions, which would be responsible for selecting the most accurate response according to the context. To sum up, this paper shows some ways in which data can be visualized and some means to analyse and interpret data resulting from the application of the DISTSEM method.

Palavras-chave : Semantic distance judgment; Stroke; Semantic network; Semantic memory.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )