A systematic review of 39 studies on brain-behaviour associations in psychiatric disorders found that canonical correlation analysis (CCA) and partial least squares (PLS) are the most commonly used methods to investigate these associations. These methods aim to capture shared information between brain and behaviour in the form of latent variables. The studies focused on associations between brain morphology, resting-state functional connectivity, and fractional anisotropy with symptoms and cognition in four diagnostic groups: attention deficit and hyperactivity disorder (ADHD), autism spectrum disorders (ASD), major depressive disorder (MDD), and psychosis spectrum disorders (PSD), as well as a transdiagnostic group (TD). Most studies (67%) used CCA and found that clinical/cognitive symptoms were linked to frontal morphology/brain activity and white matter association fibres. Other behavioural variables, such as physical health and clinical history, were also identified as important features. However, many studies were at risk of bias due to small sample sizes and in-sample testing. The review highlights the importance of out-of-sample testing to avoid overfitting and ensure generalisability. The results suggest that frontal regions may be a transdiagnostic marker for psychopathology and cognition in psychiatric disorders. The review also emphasizes the need for careful consideration of the choice and quality of behavioural measures, as well as the importance of expanding brain-behaviour studies to include transdiagnostic factors of psychiatric illness. The findings suggest that the overall effect sizes of 'doubly multivariate' brain-behaviour associations may be optimistic due to the lack of rigorous out-of-sample testing and suboptimal sample-to-feature ratios. The review concludes that 'doubly multivariate' methods are promising for investigating brain-behaviour interactions in psychiatric disorders, but further research is needed to address the challenges of overfitting and ensure the reliability and validity of findings.A systematic review of 39 studies on brain-behaviour associations in psychiatric disorders found that canonical correlation analysis (CCA) and partial least squares (PLS) are the most commonly used methods to investigate these associations. These methods aim to capture shared information between brain and behaviour in the form of latent variables. The studies focused on associations between brain morphology, resting-state functional connectivity, and fractional anisotropy with symptoms and cognition in four diagnostic groups: attention deficit and hyperactivity disorder (ADHD), autism spectrum disorders (ASD), major depressive disorder (MDD), and psychosis spectrum disorders (PSD), as well as a transdiagnostic group (TD). Most studies (67%) used CCA and found that clinical/cognitive symptoms were linked to frontal morphology/brain activity and white matter association fibres. Other behavioural variables, such as physical health and clinical history, were also identified as important features. However, many studies were at risk of bias due to small sample sizes and in-sample testing. The review highlights the importance of out-of-sample testing to avoid overfitting and ensure generalisability. The results suggest that frontal regions may be a transdiagnostic marker for psychopathology and cognition in psychiatric disorders. The review also emphasizes the need for careful consideration of the choice and quality of behavioural measures, as well as the importance of expanding brain-behaviour studies to include transdiagnostic factors of psychiatric illness. The findings suggest that the overall effect sizes of 'doubly multivariate' brain-behaviour associations may be optimistic due to the lack of rigorous out-of-sample testing and suboptimal sample-to-feature ratios. The review concludes that 'doubly multivariate' methods are promising for investigating brain-behaviour interactions in psychiatric disorders, but further research is needed to address the challenges of overfitting and ensure the reliability and validity of findings.