placeholder image

PROF Joachim Diederich


  • Artificial Intelligence (Mental Health)



Member of

  • Australian Psychological Society. Full Member 2009 -


Selected publications


Education and training

  • Habil, University of Hamburg 1995
  • PhD, Universitaet Bielefeld 1985
  • Diplom, University of Muenster 1983



Available for supervision

  • Y

Supervision Statement

  • 1. Computational methods for the assessment of mental health
    Mental health disorders are a leading cause of disability, yet diagnostic methods for the assessment of mental health problems are very subjective. There is a need for more objective techniques to diagnose major psychiatric conditions. This set of projects utilise AI techniques to determine language dysfunction in schizophrenia, depression and other disorders. The approach has been validated through a comparison with traditional forms of clinical assessment. Potential applications include the use of computerised methods for behavioural analysis to determine psychotropic drug action.
    2. Explanation and Machine Learning
    Machine learning systems such as artificial neural networks (ANNs) and support vector machines (SVMs) are commonplace in science, technology, business and health. However, the lack of an explanation capability is an impediment to the use of machine learning systems. There is no generally accepted technique that works for all systems. Also, there is no method that is tailored to the needs of a user who may not be a domain expert or have a technology background. The objective is to extract explanations from machine learning systems to support human comprehensibility.
    3. Modelling schizophrenic cognition
    Recent fMRI studies have shown that speech and language dysfunction in schizophrenia arise in part from processing abnormalities in posterior language regions. Wible et al. (2009) propose that the over-activation of the inferior parietal and posterior superior temporal regions could result in erroneous activation of semantic and speech representations, and in other symptoms of schizophrenia. Since fMRI output in isolation cannot support the inference that activated brain regions are necessary for mental processes, computational modelling plays an important role. The objective is the development of simulation mo