This thesis investigated two established neurophysiological biomarkers of schizophrenia: mismatch negativity (MMN) and P3a; purported indices of the deviance detection and orienting response, respectively. In light of recent interest into a shared diathesis model between schizophrenia- and affective-spectrum disorders, this thesis evaluated the utility of MMN/P3a in informing the underlying neurobiology of early stages of schizophrenia- and affective-spectrum disorders. Despite differences in the composition of schizophrenia- and affective-spectrum subgroups, the first two studies showed that MMN/P3a was similarly impaired in the two diagnostic spectrums, with the schizophrenia-spectrum displaying more severe/widespread impairments. The third study utilised a data-driven method and determined three clusters with distinct patterns of MMN/P3a amplitudes among patients with emerging schizophrenia- and affective-spectrum disorders. Critically, about half of the cluster with the greatest neurophysiological impairments consisted of schizophrenia-spectrum patients. The last study explored the stability of MMN/P3a over time and its value in predicting functional outcomes in schizophrenia- and affective-spectrum patients. This study showed that MMN impairments worsen over time, suggestive of ongoing pathophysiological processes. Additionally, greater deficits in MMN amplitudes at baseline were associated with the most severe levels of later disability. In conclusion, this work has been a substantial contribution to the somewhat limited literature on MMN and P3a in early psychotic (and related) disorders. Critically, this thesis supports a re-conceptualisation of the proposed neurophysiological biomarkers of schizophrenia by demonstrating a broader application of MMN and P3a (in early schizophrenia- and affective-spectrum disorders). Accordingly, these studies also provide neurophysiological evidence of a shared diathesis between schizophrenia- and bipolar-spectrum disorders.