Neural Conditioning Using Real-time fMRI

John O'Doherty, D.Phil

California Institute of Technology

Funded in June, 2005: $100000 for 2 years
LAY SUMMARY . ABSTRACT . HYPOTHESIS . FINDINGS . SELECTED PUBLICATIONS .

ABSTRACT

back to top

Neural Conditioning Using Real-time fMRI

In instrumental conditioning, an animal or human can learn to change the probability of making a particular response if that response is associated with rewarding or punishing feedback. Up until now, instrumental conditioning has been concerned with learning of overt behavioral responses. Here we propose a more radical line of investigation. Rather than conditioning overt behavior, we will attempt to directly condition neural activity. This involves training subjects to activate a part of their brain in order to obtain reward. To achieve this we will use functional magnetic resonance imaging (fMRI) to measure neural responses in real-time. A contingency will be established such that if following presentation of particular cue, a subject produces activity in a target brain area that exceeds a given threshold, a liquid juice reward will be delivered. We will focus on only two brain regions: primary motor cortex, and primary visual cortex.

We aim to show that over the course of training, subjects can learn to increase activity in these brain regions. To achieve this, we will uses shaping procedures, in which we will gradually raise the activity threshold for reward, so that over the course of training subjects must learn to continually increase activity in the target region in order to keep obtaining reward. Neural conditioning could be applied to the management or treatment of a wide variety of neurological or psychiatric disorders, including stroke, epilepsy, chronic pain, or mood disorders.

HYPOTHESIS

back to top

Hypothesis:
By using real-time event-related functional MRI (fMRI) we aim to test the hypothesis that it will be possible to train human subjects to activate parts of their brain on demand in order to obtain reward.

Goals:
Specific goals are to establish whether neural activity can be conditioned in one specific brain area: the primary motor cortex. By doing a systematic investigation of whether it is possible to condition this brain region we will be able to establish whether 1) Subjects can learn to use explicit imagining strategies in order to activate motor cortex to obtain reward.  2) We will attempt to demonstrate that it is possible to use ‘shaping’ and successive approximation procedures to increase neural activity in this area, in much the same way that instrumental conditioning procedures can be used to train animals or humans to learn complex behaviors. 3) We will determine whether it is possible to condition subjects to activate parts of their motor cortex without instructing them to use a particular explicit strategy, by making use of intrinsic variability in activity in those regions to begin the shaping procedure.

Methods:
We will use functional Magnetic Resonance Imaging (fMRI) to acquire images of volunteers' brains while they perform an instrumental conditioning task. We will measure activity in motor cortex in real-time. This involves feeding the brain images to a computer as soon as they are recorded, and then rapidly processing the images on a scan by scan basis in order to retrieve an estimate of the activity in motor cortex on that particular trial. If activity exceeds a specific threshold, we will then trigger a reward (e.g. 0.7ml of a pleasant tasting juice). Using this feedback procedure, we aim to train subjects to increase activity in their motor cortex in order to obtain reward.

FINDINGS

back to top

Lay Results
In instrumental conditioning, an animal or human can learn to change the probability of making a particular response if that response is associated with rewarding or punishing feedback. Up until now, instrumental conditioning has been concerned with learning of overt behavioral responses. Here we proposed a more radical line of investigation. Rather than conditioning overt behavior, we attempted to directly condition neural activity. This involved training human volunteers to activate a part of their brain in order to obtain reward. To achieve this we used functional magnetic resonance imaging (fMRI) to measure brain activity in real-time while volunteers were lying in the scanner. On occasion, volunteers would see a particular image, and if on seeing that image they successfully increased neural activity in the part of brain concerned with generating hand movements (but without making any real hand movements), then the volunteers would receive a monetary reward. In response to another image, subjects had to instead imagine making a foot movement in order to get the reward. Once subjects had learned to activate the correct area of their brain, we then over time increased the level of activity needed to obtain reward . This had the effect that subjects learned to progressively increase activity in specific brain regions in order to keep attaining the rewards. This neural conditioning procedure may have important therapeutic implications in the future, as it could be used to help modulate neural activity in parts of the brain implicated in particular neurological or psychiatric diseases.

Scientific Results 
Successful learning is often contingent on feedback. In instrumental conditioning, an animal or human learns to perform specific responses to obtain reward. Instrumental conditioning is often used by behavioral psychologists to train an animal (or human) to produce a desired behavior. Shaping involves reinforcing those behaviors, which in a stepwise manner are successively closer to the desired behavior until the desired behavior is reached. In this research funded by the Dana Foundation we aimed to extend this traditional approach to directly shape neural activity instead of overt behavior. To achieve this, we scanned human subjects with functional magnetic resonance imaging and performed image processing in parallel with acquisition. We delineated regions of interest (ROIs) in finger and toe motor/somatosensory regions and used an instrumental shaping procedure to induce a regionally specific increase in activity by providing an explicit monetary reward to reinforce neural activity in the target areas. After training, we found a significant and regionally specific increase in activity in the ROI being rewarded (finger or toe) and a decrease in activity in the nonrewarded region. This demonstrates that instrumental conditioning procedures can be used to directly shape neural activity, even without the production of an overt behavioral response. This procedure offers an important alternative to traditional biofeedback-based approaches and may be useful in the development of future therapies for stroke and other brain disorders.

 

SELECTED PUBLICATIONS

back to top

Bray S., Shimojo S., and O’Doherty J.P. Direct instrumental conditioning of neural activity using functional magnetic resonance imaging-derived reward feedback. J Neurosci. 2007 Jul 11;27(28):7498-507.