The present study investigated the causal predominance of cognition on anxiety, depression, paranoia, phobia and somatic concern over three time waves of self reported data measured every six months over one year, of 145 cases experiencing their first episodes of psychosis. In turn the symptoms of anxiety, depression, paranoia, phobia and somatic concern were examined for their cross-influential effects on cognition. Cognition was examined under a causal predominance hypothesis as the lead symptom because of its influence recognised in the literature under the neurodevelopmental hypothesis. These longitudinal effects were examined using structural equation modelling. Prior to this investigation, the research was able to demonstrate a stable 6-factor measurement model with these symptoms between two independent samples of early psychosis cases that met guidelines of treatment under the Australian national early psychosis treatment guidelines. This measurement model demonstrated good internal reliability and construct validity. Most symptoms over each time wave had a "domino effect" where the symptom prior to the next wave of assessment had an influence. This is known as a mediation effect. Somatic concern and depression demonstrated a "snow ball" or direct effect where the extent of the condition at time one influenced directly the condition at time three. Structural models, which examined the cross-influential effect between cognition and the other symptoms, demonstrated an effect between paranoia and cognition. This effect demonstrated that paranoia at Time 2 (i.e., 6 months after stabilisation of symptoms), had a crossinfluential effect on cognition at Time 3 (ie, 12 months after stabilisation of symptoms). / It was argued that poor thinking styles that lead to distortion in feelings of mistrust evident in the paranoia symptom, in turn led to deterioration in cognition. Other symptoms did not demonstrate a cross influential effect. Previous research suggesting that symptoms act independently of each other over time supports the results of independence of the other symptoms. Further research was suggested by linking different levels of psychosis research of the aetiological factors (e.g. genetic factors), neuropathology (e.g., reduced synapse density) and phenomenology (e.g., positive and negative symptoms) into an integrative framework. It was suggested that structural equation modelling as exemplified in the thesis could be used as a technique to examine how these differing levels could be investigated under a unified theory of psychosis based upon the neurodevelopmental hypothesis.