The problem addressed in ONOFF ERC AdG project is the underlying neuronal and cognitive mechanisms for auditory verbal hallucinations(AVH) in schizophrenia, using behavioral and brain imaging methods. A subsidiary problem is how to understand the spontaneous fluctuations over time of hallucinatory episodes, and in particular what causes the "voices" to temporarily go away, with a long-term goal of contributing to development of new interventions, targeted on a symptom rather than on a diagnosis as such. Schizophrenia is one of the most severe mental disorders, which affects about 1% of the European population, and with enormous costs for the society. There is a changing demographic pattern of the incidence of schizophrenia which goes together with the increasing urbanization and migration into the major European cities. Thus, understanding the the most severe symptom, AVH, in one of the most severe mental disorders, schizophrenia, is of major societal importance. The overall objective of the project follows a model called "Levels of explanation", which seeks to understand AVH at different explanatory levels, from the clinical to the neuronal levels. A major issue in the ONOFF project is the fluctuations of hallucinatory episodes, and if these are related to changes in excitatory/inhibitory influences at the level of neurochemistry in the brain. Using MR spectroscopy (MRS), our group was the first to report increased levels of the excitatory neurotransmitter Glutamate in brain regions which are activated during AVH episodes. We have recently followed-up these initial results, showing that frequency and severity of auditory hallucinations correlate positively with increased levels of glutamate in temporal (STG) brain regions, but negatively in frontal (ACC) regions, as would be predicted from the VOICE model proposed by Hugdahl in 2009. In order to relate these findings to findings at the level of neuroimaging, we have developed a new MR sequence to simultaneously assessing brain transmitters and functional changes seen in fMRI BOLD data. Preliminary validations show that the simultaneously acquired MRS and BOLD data is a feasible way forward. As stated in the proposal, there is a need for new approaches to cognitive training of auditory hallucinations, as well as new ways of acquiring data on frequency and content of auditory hallucinations in real-time. We have for this purpose developed two smartphone apps, one for training, and another for symptom capture screening
