Specificity of Working Memory difficulties in children with NLI

A large number of studies have focused on the investigation of phonological working memory (PWM) confirming that children with SLI show difficulties in storage and processing of verbal information (e.g. Gathercole & Baddeley, 1990; Conti-Ramsden & Durkin, 2007). Visuo-spatial WM skills have only recently gained researcher’s interest. For the defenders of the general cognitive capacity limitations theory, the specific nature of the material (verbal vs visuo-spatial) is less important than how this material is mentally manipulated (Ellis-Weismer, 1996; Johnston, 1991, 1994). In that perspective, there is no reason why we should not assume that children with language impairments would present difficulties in storage and processing of visuo-spatial material. To date we have only few studies investigating visuo-spatial WM in children with SLI. The findings in these studies are contradictory, perhaps due to their methodological differences (e.g. immediate vs differed recall tasks, serial vs simultaneous presentation of stimuli etc.), and therefore non-conclusive. For Archibald and Gathercole (2007) the poor performance of children with SLI on complex memory tasks involving visuo-spatial processing and verbal storage, depict a deficit located most possibly on the central executive of Baddeley and Hitch’s tripartite model (1974) and not on the visuo-spatial sketchpad. For Parisse and Mollier (2008), poor performance in visuo-spatial tasks in children with SLI are only found in the case of serial presentation of stimuli demanding more elaborated encoding skills (e.g. Corsi’s blocks) than simultaneous presentation (e.g. visual patterns). Lastly, for Hoffman and Gillam (2004) poor performances of children with SLI on both complex visuo-spatial (processing and storage) and verbal memory tasks (processing and storage), prove that children with SLI show deficits not only on phonological WM but on the visuo-spatial sketchpad as well.

The aim of our study (chapter 5) was to investigate visuospatial WM skills in children with NLI, presenting both verbal and nonverbal difficulties. We compared a group of children with NLI (N=15) to a group of lexical-matched younger TD children (N=15) and a group of low-IQ children (N=15). The material consists of two experimental tasks inspired from the Peanut-task (de Ribaupierre et al. 2000), an immediate recall task (simultaneous presentation of stimuli) and a recognition task, demanding different levels of encoding skills (the recognition task being easier than the recall task). The findings from the recall task suggest that NLI children present a developmental delay when asked to restore (the order is not important) a series of visuo-spatial stimuli (from 3 to 5) in either close or dispersed arrangement. Their performances on this task are similar to those of the TD group. On the same task, children with NLI show significantly better performance than low-IQ children, due to their difference on nonverbal IQ. The simultaneous presentation of stimuli, like in the study of Hick et al. (2005) allows children to process stimuli globally. Although, this involves an easier encoding than a serial presentation task (Parisse & Mollier, 2008), we found that children with NLI show limitations in their visuo-spatial sketchpad (Table 22). In the recognition task we used various experimental conditions, differing in the degree of difficulty of the encoding process. In particular, some of the conditions required ‘counting’ of the stimuli to verify that their number is correct (e.g. identical and wrong number conditions) whereas others did not (e.g. wrong place condition) involving more global perception of stimuli. Performance of children with NLI was similar to those of younger TD children. Moreover, the two groups presented similar profiles in all conditions suggesting a developmental delay (Table 23 and Figure 25).

Further research is needed to investigate storage and processing of visuo-spatial material especially by comparing children with NLI to children with SLI in order to study if the two groups present similar profiles. If not, then we could assume that the poor performances of children with NLI in this task reflect important limitations in general cognitive capacity and use this aspect to differentiate the two groups. However, it is worth mentioning that we obtained a significant difference between the NLI and the low-IQ group (matched in age, verbal IQ and reading age), the latter performing significantly lower than the NLI group in both tasks, probably because of their lower nonverbal IQ. Pickering (2004) evokes four possible mechanisms, which could be related to the development of the visuo-spatial memory: knowledge increase, strategy use, processing speed and attention capacity. Thus, the difficulties of children with NLI in this study could reflect associated cognitive difficulties, for example in attention, aspect we did not evaluate during this study. Further investigation is, therefore, needed.