The Promise of Neurofeedback

by Carl Sherman

September 10, 2013

Psychiatric illness is brain disease. Decades of neuroimaging, molecular biology, and other research have given us a picture of  what seems to go wrong—overactive circuits in anxiety, missed connections in schizophrenia, dysregulated neurotransmitters in depression—but  gaps in our knowledge remain frustrating.

The gap between brain biology and psychiatric treatment is even wider.  Drugs produce sledgehammer alterations in brain chemistry that relieve symptoms in some people, sometimes, often at a heavy cost of adverse effects.  Brain stimulation can be effective, but we don’t yet know why. Psychotherapy apparently changes brain function, when it works, but just how also is veiled in mystery.

 If the science to support psychiatry isn’t there yet, it’s not for lack of trying, as researchers struggle to penetrate the brain's secrets to design better treatments for its dysfunctions. 

One of the most promising paths toward this end combines a powerful tool of basic research, functional MRI, with a traditional method for altering behavior, operant conditioning.

Real-time fMRI (rtfMRI) neurofeedback shows people in an imaging scanner a continuous readout of their metabolic activity  (reflected by blood oxygenation) in a specified brain area and trains them to modify it, much as conventional biofeedback trains people to control high blood pressure or relieve headaches by altering  physiological variables like respiration rate and muscle tension.

"The neurofeedback idea is intrinsically based on the findings of neurobiology," says Nikolaus Weiskopf of University College London who was part of the research team that showed its feasibility just over a decade ago. "It ties in nicely with the general trend toward biologically-based psychiatry."

Interest in the field is building rapidly: almost as many rtfMRI neurofeedback papers (73) were published  in 2011-2012 as in the prior 10 years, and  it had its first  international conference in 2012.

In part, Weiskopt suggests, this is simply the expected growth curve of an exciting new technology, as word gets out and such practical barriers as funding and buying equipment are overcome. "But it also reflects the generation of new ideas; as more people come into the field, there are more kinds of thinking."

RtfMRI was developed as a tool for basic brain research,  but "clinically-related applications came up rather quickly, which is different from other technologies," Weiskopf says. "That's why we went into the field in 2000: with a long-term view of applying it to patients."

Training down anxiety and depression

 While he and other researchers emphasize the distance between encouraging findings in the lab and real help for real people, there's no question that recent work is indeed encouraging.

In a study reported in the April 2013 Translational Psychiatry, Michelle Hampson and colleagues at Yale used neurofeedback to train participants—people who reported high levels of contamination anxiety— to modify activity in the orbitofrontal cortex (OFC), a brain area believed to play a key role in emotional regulation and to function abnormally in psychiatric  symptoms like anxiety.

Two sessions in the scanner reduced their anxiety response to provocative images, and produced alterations in brain function that went well beyond the target area.

"The OFC is part of a big network," Hampson said. "We looked at resting state connectivity, which reflects functional brain organization." Their findings included reduced connectivity to regions associated with emotional processing, and increased connectivity with prefrontal regions tied to cognitive control.

Both brain and behavior changes persisted when tested after the training period (an average of three days later)—one of the few fMRI neurofeedback studies to show lasting effects.

While the study involved non-patients, "potential therapeutic applications are broad," Hampson said. "The data suggests [neurofeedback] might work for obsessive compulsive patients [where contamination anxiety is a frequent symptom], and for anxiety in general.” It has implications for other psychiatric disorders as well, she says, insofar as OFC circuits often appear to be involved when people have trouble with emotional control.  

Another recent study found an effect of neurofeedback in actual patients who suffered from depression.

Here, the idea was to increase activity in brain areas associated with positive emotions: i.e. where, for each subject, the contrast was strongest between activity when viewing  positive (e.g. erotic), versus neutral or negative (threatening or disgusting) images. These target areas  varied  among   individuals (and from session to session), but by and large, "were in keeping with areas described in previous work on emotional regulation,"  says David E. J. Linden of Cardiff University, first author of the paper that reported the work in  the June 4, 2012, PLoS ONE.

The patients were instructed to increase activity in designated areas by raising the dial of a thermometer representing the fMRI signal there. "They were not given any instructions on strategy," Linden said.  Most participants accomplished the increase progressively over the course of four sessions, for the most part using personal images such as pleasant memories.

Their depression symptoms improved significantly, too; enough, in two of eight patients, to constitute remission. Similar improvements were not seen in controls who were instructed to use the same mood-elevating strategies as the test group, but without neurofeedback.

"The most important finding of this initial study was that neurofeedback is feasible in patients with depression, who often have motivational deficits," Linden said. "It was acceptable to them. In fact they were happy to come back for repeated sessions in the scanner; they seemed to enjoy the process, which can't be said for many activities they do."

The improvement in symptoms was heartening too, "but it wasn't a real efficacy trial, so we have to be cautious,"  he said.  

The road ahead

RtfMRI neurofeedback has even shown some promise for a notoriously severe and complex disorder—schizophrenia. The symptoms of this illness are extremely broad, including hallucinations, blunted emotions, and thinking distortions, and so are the brain abnormalities suggested by neuroimaging. In fact, many believe that behind schizophrenia is generalized failure of connectivity among widespread brain regions.

"Because schizophrenia is so complex, attacking it all at once to test a hypothesis in a scientific study is not possible," says Ranganatha Sitaram, of University of Florida. In an experiment reported in Human Brain Mapping in January 2013, Sitaram and colleagues focused on  the ability to recognize facial expressions, which is impaired in schizophrenia, and  targeted the insula,  a brain region associated with this faculty, particularly for negative emotions.  After four training sessions, patients were able to increase insula activity, and improve their ability to identify faces expressing disgust.

Imaging at the end of the period also suggested that training had enhanced connections between the insula and other brain areas; beyond that, the density of overall connectivity across a broad emotion-processing network increased. "If rtfMRI can be used to 'repair' neural disconnection, this could have important therapeutic implications for schizophrenia," the researchers wrote.

Before such implications for this or other disorders can be realized , studies will need "to make sure [self-regulation skills] can be generalized in real-world experience," says Nikolas Weiskopf. “Lying in a scanner is very different from everyday life.”  Effects must be shown to endure for months, rather than days, and practical questions, such as treatment length, must be clarified.

At best, "neurofeedback shouldn't be seen as a silver bullet.... [but]  integrated into a wider setting of clinical support," he says. "It might act synergistically with parallel treatments, like cognitive behavior therapy or medication."

Movement toward a therapeutic reality is perceptible, however, in the fact of registered clinical trials now in progress: David Linden is applying neurofeedback for depression to a larger group of patients; and Michelle Hampson and colleagues are testing rtfMRI for people with Tourette's syndrome, a disorder of life-disrupting tics, following basic research (supported by a Dana Foundation grant) that showed the technique could modify connectivity to a relevant brain area in healthy subjects.

If his research goes well, Linden predicts, large multicenter trials could follow in a year or two, and [if that succeeds] within five years, "clinics offering neurofeedback as a component of an integrated biopsychosocial approach to depression might be a realistic proposition."