Periodic Reporting for period 1 - DOPANF (Dopaminergic midbrain modulations by (adaptive) neurofeedback)
Reporting period: 2018-06-01 to 2020-05-31
Value-based decision-making is crucial for human behavior and learning, pervades our daily life and goes astray in psychiatric disorders such as schizophrenia and depression. The neural activity of brain regions producing and being influenced by the neurotransmitter dopamine has been associated with value-based decision-making processes. By extension, one would expect dopamine systems, mainly midbrain regions and their striatal projections, to play a role in value-based decisions, such as whether a specific reward is worth enduring some punishment. It is a matter of debate, but of surprisingly little empirical investigation, whether the rewarding and punishing aspects of value differentially or similarly depend on the dopamine system. The project DOPANF aimed to clarify this issue by investigating dynamic changes of dopamine-related neural activity. DOPANF developed two new paradigms to shed light on this important, but neglected topic for public health and subjective well-being. Specifically, DOPANF used recent neuroscientific and technological developments: real-time imaging for immediate analysis of neural activity in the midbrain and graph-theory connectivity analysis for algorithmic capture of related activity spatially distributed throughout the brain. By observing their own brain activity as neurofeedback in real time, participants learned to volitionally up- and downregulate activity in the dopamine system. The first data analysis revealed new findings about the mechanisms of learning the self-regulation of the dopaminergic midbrain. I investigated the mechanistic control of the dopaminergic midbrain. The findings underpin the theory of reinforcement learning as learning mechanism for the first time in human brain data. Furthermore, the analysis involved the spatially distributed cognitive control network in successful self-regulation of the dopaminergic midbrain. On the behavioral level, we found that learning of volitional self-regulation depends on individual sensitivity to reward. These findings might help tailor future experiments and treatments to specific participants and patients.
The second experiment on adaptive neurofeedback control of the dopaminergic midbrain is currently still exploited and will give novel insights into the process of making decisions to exert mental effort dependent on dopaminergic activity. With this, DOPANF aims to achieve a more precise understanding of decision-making and elucidate the exact role of the dopaminergic midbrain in value processing and its impact on effort discounting. Overall, these findings provide a knowledge basis for a better understanding of psychiatric disorders involving the dopaminergic midbrain, such as depression, schizophrenia and addiction.
The second experiment on adaptive neurofeedback control of the dopaminergic midbrain is currently still exploited and will give novel insights into the process of making decisions to exert mental effort dependent on dopaminergic activity. With this, DOPANF aims to achieve a more precise understanding of decision-making and elucidate the exact role of the dopaminergic midbrain in value processing and its impact on effort discounting. Overall, these findings provide a knowledge basis for a better understanding of psychiatric disorders involving the dopaminergic midbrain, such as depression, schizophrenia and addiction.
In summary, the work on project DOPANF comprised two actions about neuroscientific experiments using real-time fMRI neurofeedback investigating reward and punishment processing in human brains. Beyond the scientific questions about these dopaminergic pathways, it comprised the development of a novel paradigm and new cutting-edge algorithms for data analysis. Furthermore, with the project DOPANF I established a real-time fMRI setup in the SNS lab, which is freely available to other researchers in the lab. Furthermore, I continued to develop an open-source software with two new cutting-edge algorithms.
To achieve the scientific objectives, several efforts in software development have been necessary, which are available as open-source tools for the public:
1) A real-time fMRI neurofeedback toolbox for analysis of the data at the MR facility:
https://econgit.uzh.ch/rt_heroes/rt_fmri_opennft_experiments
2) A behavioral paradigm on decision-making dependent on individual’s brain activity
https://econgit.uzh.ch/ninwie/task_mentaleffortdiscounting_nback
3) A novel analysis approach using graph-theory based analysis for connectivity analysis of real-time fMRI data
https://github.com/lydiatgit/Connectivity_SINGLE
4) A toolbox for data quality assurance of real-time fMRI data (collaboration project)
https://github.com/rtQC-group/rtQC
The exploitation and dissemination of DOPANF in summary:
I attended four conferences and two workshops presenting the project DOPANF.
Two manuscripts as first author are currently under review and preprints are already available. Several other manuscripts I contributed to are also in preparation by colleagues.
I supverised one master project thesis dealing with a novel data analysis approach for DOPANF fMRI data.
I organized or have been invited to four different outreach activities to present my research to the general public.
Two press releases and one YouTube video have been created presenting the project DOPANF.
To achieve the scientific objectives, several efforts in software development have been necessary, which are available as open-source tools for the public:
1) A real-time fMRI neurofeedback toolbox for analysis of the data at the MR facility:
https://econgit.uzh.ch/rt_heroes/rt_fmri_opennft_experiments
2) A behavioral paradigm on decision-making dependent on individual’s brain activity
https://econgit.uzh.ch/ninwie/task_mentaleffortdiscounting_nback
3) A novel analysis approach using graph-theory based analysis for connectivity analysis of real-time fMRI data
https://github.com/lydiatgit/Connectivity_SINGLE
4) A toolbox for data quality assurance of real-time fMRI data (collaboration project)
https://github.com/rtQC-group/rtQC
The exploitation and dissemination of DOPANF in summary:
I attended four conferences and two workshops presenting the project DOPANF.
Two manuscripts as first author are currently under review and preprints are already available. Several other manuscripts I contributed to are also in preparation by colleagues.
I supverised one master project thesis dealing with a novel data analysis approach for DOPANF fMRI data.
I organized or have been invited to four different outreach activities to present my research to the general public.
Two press releases and one YouTube video have been created presenting the project DOPANF.
Beyond the state-of-the-art, DOPANF indicates reinforcement learning in the dopaminergic midbrain as the learning mechanism behind neurofeedback training. This was only been described in theoretic models before. Furthermore, the adaptive neurofeedback based on the neural activity from the dopaminergic midbrain leads to a individual approach of intervention with human brains. This innovative feature from project DOPANF is the possibility of subject specific therapies. Such a Personalized Medicine is a special topic in the EU Health programme, which notes that many of our most common medicines are not effective in treating large numbers of patients. DOPANF established an approach of mental health training dependent on the individual’s neural activity. The findings of project DOPANF are highly relevant for clinical use in schizophrenia, depression and addiction diseases. The established mental training and individual decision-making approach might help to detect biomarkers for psychiatric diseases, such as depression and addiction.