Learn to Control Your Own Brain Emotion Networks w/ Real Time fMRI Neurofeedback

Neuro-researchers in the UK and the Netherlands helped establish real-time fMRI neurofeedback as a practical tool for self-controlling emotional networks in healthy subjects, back in 2009 (PDF). Since then, they have studied the use of real time fMRI feedback in the treatment of depression, with some positive results. Here is more from their most recent study:

a) Activation of the insular cortex (INS) bilaterally and the right ventral striatum (VS) supported the neurofeedback task, whereas the temporoparietal junctions (TPJ) of both hemispheres were deactivated. The TPJ is recognised as part of the brain’s “default mode network” that is deactivated during effortful tasks.
plosone

Depression is the mental disorder with the largest impact on public health. Up to 20% of the population suffers from a depressive episode at some point in their lives [1], and major depressive disorder (MDD) is a main source of disability for adults of working age in industrialized countries. At least 30% of patients with MDD do not respond to standard pharmacological and/or psychological treatments [2], and a considerable number of those who do respond initially go on to develop a chronic relapsing-remitting disorder. These patients with no or only a partial response to standard treatments often enter a vicious circle of psychosocial decline with further deterioration of their mood and level of functioning. To prevent relapses new therapeutic strategies have to be developed that aid the restructuring of cognitive schemas and might even prevent the formation and crystallization of dysfunctional thought patterns during early phases of depression.

...In the present study we localized areas responsive to positively valenced visual stimuli adapted from the International Affective Pictures System (IAPS) [15], [10] and then trained patients with unipolar depression to upregulate the activity in this target region over four sessions. We hypothesized that the combination of the physiological upregulation and the reinforced training of positive thought patterns would lead to an improvement of mood, which would not be seen in a control group that engaged in an emotion regulation protocol without neurofeedback. _PLoS ONE

Higher activation of right insula (INS), ventral striatum (VS), anterior cingulate cortex (ACC) and ventromedial prefrontal cortex (VMPFC) during presentation of positive compared to neutral images in the localiser runs (for full list of areas see Table 2). The localiser runs were effective in identifying brain areas responsive to positive images, which were used as target regions of interests (ROIs) for the subsequent neurofeedback procedure.
plosone

Patients in the NF group were trained to upregulate brain areas responsive to positive emotions using a procedure modeled on our previous work with healthy participants [10]. A target area was identified by the contrast between responses to positive and neutral images in a localizer scan to ensure that an area involved in positive emotion processing was selected. In the localizer scan, we assessed brain responses to positive, negative and neutral pictures by presenting four pictures of the same emotion category in blocks of 6 s (1.5 s per picture), alternating with a fixation baseline of 12 s. We presented 12 blocks per category in pseudorandom order. We used pictures from the IAPS [15] with negative (mean normative ratings for valence 2.8 [SD.42], arousal 5.63 [SD.55]), positive (valence 6.90 [.55], arousal 6.00 [.74]) and neutral valence (valence 5.45 [.56], arousal 3.44 [.47]). Pictures showed, for example, scenes of danger or disgust in the negative category, and scenes of romance including mild erotica or exciting sports in the positive category. After the localizer scan, patients were trained to upregulate the target area during three neurofeedback scans lasting ca. 7 minutes each per session (Fig. 1). Patients were informed about the general function of the target area but were not given any specific instructions about strategy. The task we set for them was to increase activity in the target area by as much and as consistently as possible.

...For the neurofeedback, a continuous signal from the target area (updated every TR and thus every 2 seconds) was displayed using the picture of a thermometer whose dial indicated the amplitude of the fMRI signal in the target area. Changes in the amplitude were indicated as the percent of signal change, calculated using the current signal intensity value and comparing it with the average value determined from the rest period immediately preceding each upregulation block. The scaling of the thermometer was in steps of 0.05%, with a maximum value of 0.5% (see Fig. 1). A change of background colour every 20 s indicated to participants whether their task was to regulate (green background) or rest (yellow background). The online GLM was computed with one predictor for the regulation state, convolved with a haemodynamic reference function. The top one-third (defined by the t value for the contrast between the regulation predictor and baseline) of the voxels from the target region was used to compute the feedback signal. For runs in which participants failed to upregulate the target area during the regulation periods (negative percent signal change), another target area was selected for the next run, using the cluster with the strongest activation for the regulation predictor. This adjustment in the target area was necessary in 15/32 (47.9%) of the sessions after the first NF run, and in 4 sessions after the second run. The reasons for this approach were two-fold. First, the adjustment of ROIs aided the shaping of mental strategies in the desired direction. Shaping is a common concept in the operant learning of a highly demanding task [11]. Secondly, our focus was not so much on the ability of participants to learn to regulate a specific brain region but on the effects of the NF training procedure on participants’ mood.

...Patients in the NF group reported initially using imagery of the positive scenes in the localizer scan in an attempt to increase activation in the target brain areas, but they later changed to evoking memories and imagery of autobiographically relevant material. For example, the happy memories that they reported as successful strategies included holidays, thoughts about their family being happy, and imagery of beautiful scenes from nature. Some patients attained good self-regulation of the target areas through mental simulation of future successes, and one patient successfully used imagery of an out-of-body experience. Conversely, during rest periods, the patients reported trying to “empty their thoughts” and to meditate. Patients in the IM group were instructed to engage in similar strategies as those reported by the NF patients. At debriefing, they confirmed that they had used these strategies. No patient reported any distress arising from the procedure.

...In the present study, four sessions of non-invasive fMRI-neurofeedback reduced the symptoms of depression with an effect size similar to those obtained with deep brain stimulation (DBS) [3]. Although the mental strategies of positive thoughts, memories, and imagery may have played a considerable part in this improvement, the neurofeedback procedure was crucial as evidenced by the absence of any clinical improvement in the control group. _PLoS ONE

More study details at the link above.

The researchers intend to expand their research in future studies so as to deal with possible confounders, and to introduce increased rigour via randomisation and more sophisticated control procedures.

There are many advantages in the ability to detect and shift the activity in one's own emotion networks. This could be true both in one's professional and personal life. It is also likely that the ability to control specific brain networks will facilitate learning -- particularly in difficult subjects where frustration and apprehension can be a factor.

fMRI neurofeedback is a bit unwieldy, given the bulk and expense of fMRI scanners. Sophisticated EEG neurofeedback can achieve essentially the same results with less expense, although the increased neuroscientific rigour of visualising fMRI network activation and deactivation, is an advantage in the research setting.

Once protocols are developed with fMRI, then parallel protocols using advanced -- but portable -- EEG can be cross-validated using data from the fMRI.

The microprocessor revolution has helped to shrink the size and reduce the weight of a wide range of sophisticated electronic devices. Home EEG equipment is already available which can be used in conjunction with a pad computer or a smart phone. As these devices improve -- and as the protocols for self-control of brain networks continue to be developed and improved -- expect to be able to teach your brain to control itself in more ways than you might currently imagine.