Search

Executive function (EF) is an umbrella term for a group of related yet distinguishable cognitive functions responsible for flexibly controlling and directing attention and behaviour. EF is important in the top-down control of decision-making and the ability to direct attention towards the fulfilment of long-term goals. EF, particularly in emotive contexts, is still developing during adolescence. EF may also be affected negatively by various threats to optimal EF development, including depression and HIV infection. Immature or suboptimal EF may predispose to risk-taking behaviour, including risky sexual behaviour and substance use, which in turn increases the risk of HIV transmission and poorer health and socioeconomic outcomes. I have three major hypotheses which are explored in this thesis. Firstly, I hypothesise that there is a relationship between HIV and poorer EF in adolescents. Secondly, I hypothesise that poorer EF is associated with increased risky sexual behaviour and substance use in adolescents, and lower adherence to treatment in HIV-positive adolescents. Thirdly, I propose that musical training, particularly group drumming, is a feasible and acceptable intervention to enhance EF and mood in HIV-positive adolescents. The thesis begins with an introduction to the relevant concepts and background literature (Chapter 1) before presenting a systematic review and meta-analyses of EF in HIV-affected children and adolescents globally (Chapter 2). It then shifts to the region of interest, rural South Africa, and presents the validation of a tablet-based EF assessment tool (Chapter 3) followed by a cross-sectional study assessing the relationship between EF and risk-taking behavioural outcomes, including behaviourally acquired HIV, in late adolescent females (Chapter 4). Finally, musical training is explored as a potential intervention to improve EF. A literature review on musical training and EF, and the potential socio-emotional effects of group drumming (Chapter 5) provides the background to a feasibility trial assessing the feasibility and acceptability of a group drumming intervention aiming to improve EF in HIV-positive adolescents in rural South Africa (Chapter 6). The thesis concludes with a general discussion chapter which integrates the findings and considers potential future directions (Chapter 7). Study 1 found evidence for EF deficits in perinatally acquired HIV; evidence was limited for behaviourally acquired HIV. Study 2 found that the OCS-EF is a valid and feasible tool for measuring EF reliably amongst adolescents in rural South Africa. Study 3 found that EF is only weakly associated with certain risky behavioural outcomes, and that adolescents with behaviourally acquired HIV may only have limited EF deficits relating to verbal working memory. Study 4 found that a group drumming programme is a feasible and acceptable intervention amongst HIV-positive adolescents in rural South Africa, with promising effects on psychosocial wellbeing, including mood, but no effects on EF.

The experiments in this thesis explored the development of executive function in 5- to 11-year-old children. Developmentally-appropriate versions of the task-switching paradigm, go/no-go task and self-ordered pointing test were used to measure shifting, inhibition and working memory respectively. These executive skills were examined independently and within-task experimental manipulations were used to explore both the executive and non-executive processes that influenced children’s performance. These allowed the investigation of not only when, but also how executive function develops. It was found that shifting development, as measured by the task-switching paradigm was highly influenced by the specific tasks switched between and the conflict created by the overlap of the tasks, as well as by previous task experience. Working memory for pictures was also influenced by previous experience and task difficulty, however the predicted relationship between memory for nameable objects and language ability was not found. Inhibition on the go/no-go paradigm appeared to be driven by an improvement in the efficiency of response inhibition enabling older children to inhibit a response at an earlier stage during the movement. Shifting, inhibition and working memory all showed developmental improvements during mid-childhood, demonstrating the protracted development of executive function. Shifting and working memory showed a similar pattern of development whereas inhibition reached a stable level of performance at an earlier age. There were no correlations between the three executive skills studied in this thesis, supporting the fractionation of executive function.

From a literature review, it was concluded that schizophrenia primarily involves deficits in conscious, controlled processing. This was shown to be compatible with Frith's (1992) neuropsychological model of impaired meta-representation in the disorder. There is strong evidence that patients with paranoid symptoms or behavioural signs have deficits in the representation of others' mental states ('theory of mind'; ToM), but two recent studies have produced conflicting results. Those findings were reconciled in the first study of the thesis, which showed that, on false belief tasks, patients have intact first-order ToM, but specific impairments at the second-order level. This was later confirmed using a 'hints' test of ToM. The results were contrasted with the case of autism. On a spatial reversal test of executive function, schizophrenics with behavioural signs made more perseverative errors than controls. No correlations appeared between ToM and executive function for any of the schizophrenic symptom groups. This was contrasted with the case of autism, and it was suggested that schizophrenia involves late-occurring, independent deficits in separate metarepresentational domains. It was argued that Frith's model of schizophrenia can be extended to include impaired representation of own knowledge, explaining the deficient use of context in the disorder. Some evidence was obtained that patients with primarily behavioural signs are impaired at naming objects in a picture context; this ability was unrelated to ToM, consistent with independent deficits in separate metarepresentational domains. It was suggested that 'weak central coherence' in autism may also reflect impaired representation of own knowledge, and some evidence was obtained that (like autistics), symptomatic schizophrenics show facilitation on embedded figures tests. Schizophrenic patients performed the same as controls, however, on an illusions task. For patients with behavioural signs, embedded figures accuracy was inversely related to the ability to name objects in a picture context, and it was argued that this supported task analyses suggesting a common cognitive process.

Three separate studies employed colour tasks to investigate aspects of cognitive impairment in AD. In the first study an information-processing framework was developed using a model of colour processing (Davidoff, 1991, 1997), and the framework guided the design of a number of colour tasks. Contributions of AD and impaired colour vision to task performance were considered, and findings suggested impaired memory and executive functions but relatively intact perceptual functions in AD. Executive functions were the focus of the second study. AD patients are typically impaired on the form sorting component of the Colour-Form Sorting Task (Weigl, 1941) and this impairment was hypothesised to originate from an executive function deficit. The decision-making load of the CFST was varied, and findings suggested that impaired form sorting may originate from a deficit in selective attention to competing stimulus dimensions. The third study was a case study of a 61 year-old woman with visual variant AD who presented with two specific deficits: global processing impairment and inability to name the colour of the ink in the Stroop task. Approaches to the study of selective attention can be divided into two types: holistic and analytic (Shalev & Algom, 2000). The holistic approach was used to investigate her global processing impairment, and the impairment appeared to originate from a feature integration deficit. The analytic approach was used to investigate her Stroop impairment. Findings suggested a deficit in selective attention to competing stimulus dimensions that was perhaps more severe than that shown in standard AD; letter form/colour are for most people separable dimensions, but they appeared integral for her. Participants found the tasks non-anxiety provoking and enjoyable to complete, therefore future research might aim to develop a more comprehensive battery of colour tasks to assess cognitive impairments in AD.

The work presented in the thesis is based on the assumption that two types of cognitive processes can be differentiated, domain-specific (i.e. modular) and domain-general (i.e. central) processes. Autism and dyslexia are viewed as disorders with domain-specific deficits, while mental retardation is viewed as a disorder in central processing. Using Anderson's (1986) inspection time paradigm it was shown that central processing speed was intact in autism and dyslexia but impaired in mental retardation. The modularity status of phonological processing was investigated using the Phonological Assessment Battery. It was found that test performances did not only require intact phonological skills, but also intact domain-general functioning. The interaction between modular and central processes was studied by examining surface form retention levels for scrambled and ordered stories. The results showed that the autistic group had enhanced, the dyslexic and MLD groups suppressed surface form retention levels. It was further found that half of the subjects in the autistic group were hyperlexic, and that these subjects, in particular, represented enhanced levels of surface form. It was suggested that weak central coherence could explain this finding. Poor surface form retention in dyslexia was viewed as a secondary consequence of the phonological processing deficit. Representation of gist was examined for the same stories and in the same conditions, showing that enhanced piecemeal processing of text in the hyperlexic group was accompanied by poorer gist representation for ordered stories. The dyslexic and the MLD group were unimpaired in gist representation. These findings indicated that the abnormality in central processing in autism (i.e. weak central coherence) affected both low-level (surface form) and high-level (gist representation) central processes. In dyslexia gist representation was unaffected by modular, phonological processing deficits. The hyperlexic and non-hyperlexic subgroups were further investigated with respect to differences in the degree of weak central coherence, executive and Theory of Mind dysfunctioning. The results indicated that weak central coherence might be contributing factor in hyperlexic developments. There were no group differences in ToM, but some indication that hyperlexic subjects had greater executive function impairments.