Report on the Seville Conference of the Society for Applied Neuroscience (SAN)

by Siegfried Othmer | May 21st, 2008

It is said that an emerging democracy should be judged not by its first election, but rather by its second. Perhaps similarly a new organization should be judged not by its first conference but by its second. The Society of Applied Neuroscience just held its second biannual conference in Seville, Spain, and by the above standard seems to be well on its way. The SAN spun off from the ISNR after the Winterthur conference in February of 2004. It held its first conference in Swansea in 2006. On the order of 200 people came to this one.

One motivation for the organizational separation was the feeling among some European academics that the ISNR was not sufficiently scientifically rigorous. One could also read that to mean it was not dominated by academics. In response, the conference program certainly reflected a high level of meaty scientific content. But when it came to the neurofeedback presentations by the academics, the disconnect from the clinical world was only too apparent.

The Symposium on ADHD was dominated (unsurprisingly) by the slow-cortical-potential (SCP) work that has been incubating at Tübingen for many years. In this method, a trainee watching his own quasi-dc EEG potential is asked to produce upon command a change by several microvolts over a period of six to eight seconds, either in the direction of greater or of lesser activation (achieving bilateral control). In a couple of the presentations, comparison was made between this method and the SMR/theta training that is more traditional in the U.S. in application to ADHD.

I was distressed to find that the training parameters for the SMR/theta part of the research did not match what was routinely done in the US by the proponents of the method. The Europeans were rewarding broad bandwidths of seven and eight Hertz, using either 13-20 Hz or 12-20 Hz as the filter bandwidth. And in one comparison, the SMR/theta presentation was given in discrete epochs much like the SCP training. The more traditional approach by Lubar, Sterman, and others has been to reward a much narrower bandpass centered on 13.5 Hz, typically on the order of 3 Hz. Michael Tansey used 14 Hz as captured by the 1-Hz wide bin of his FFT program.

The use of wide bandwidths presents a very different kind of challenge than the use of more narrow filter bandwidths. In order to think through what we are asking the brain to do, I find it useful to adopt the “brain’s perspective” as much as possible as a kind of exercise. The brain organizes the construction and deconstruction of the SMR-spindles (and all other spindles), and it maintains a kind of no-man’s zone between spindles in the frequency domain. Hence the brain treats each spindle as a distinct event, one that is separate from nearby events in the domains of frequency, time, and space.

That being the case, the brain appears to relate a lot more intimately to information relating to something that it is actively managing (the individual spindle-burst) than to some statistical measure of more global activity that may have more diverse sources. In practice, matters are confounded by the fact that nearly everything works in neurofeedback to some level, but there is much to convince us that the brain is in “narrow focus” when it is organizing specific spindle-bursts, and we do better in feedback when placing the brain into narrow focus as well when this activity is to be reinforced.

On top of everything else, the results shown for SCP-type training were not very good, and in their hands the SMR/theta training wasn’t any better. There are two major problems with the SCP work to date. There is a substantial cognitive demand, so the technique is not available for very young children or for severely compromised brains. Secondly, there is a high fraction of non-responders, ranging from thirty to fifty percent. In my state of growing discomfort and distress, I blurted out at one point that I did not think the results were very good, and that if we had to live with such results clinically, we would be out of business. I regretted my outburst immediately, but too late. It had been my trauma speaking….

During the early years of our teaching neurofeedback, we always cautioned people against alpha training. It had, after all, been thoroughly discredited in research. It was Peniston’s work that got us to take a second look, despite the fact that the AAPB people tried their best to discredit him. The ham-fisted research that had gotten published during the early years totally eclipsed the wonderful work on alpha training that was being done at the Menninger group, and by people such as Les Fehmi and Jim Hardt.

And here we are again. This wimpy research will be published and recognized as the work of appropriately discerning people. It will almost surely eclipse the superb work of competent clinicians, whose work will once again be dismissed as cases of over-claiming or of selective reporting.

My fears in this regard were further reinforced when Ed Vermetten presented his invited paper next day titled “Windows of Opportunities in PTSD Neuroscience Research.” His talk covered neurofeedback as well, but the only paper he could find was the one by Graap and Freides in which they took apart Eugene Peniston after they failed to replicate his results. No mention by Vermetten of Peniston at all. Hence he was unaware of any positive potential of Neurofeedback in application to PTSD. Once again, ham-fisted research displaces legitimate findings.

At issue here is what I call the lawnmower model of neurofeedback research. The researcher starts the motorized lawnmower and sets it at the edge of the grass and expects it to cut the lawn. (This is what is meant by a double-blind study.) Now if a person actually intervenes and the lawn actually gets sensibly cut, then that person will have corrupted the pristince process, rendering the event unpublishable. Perhaps it was just good intentions all around, on the part of the therapist and the subject, and symptoms subsided. Somehow research methods cannot handle the fact that in Neurofeedback there is both a procedure involved and some intelligent guidance needed along the way, and perhaps some midcourse corrections, along with humane interaction with the client from beginning to end.

Several talks at the conference I found to be of particular interest. Thomas Elbert (University of Konstanz) presented on the current state of research on tinnitus. (It turns out that Elbert had actually been involved in the early years with Niels Birbaumer’s neurofeedback research in Tübingen.) Tinnitus appears to originate unambiguously in hearing loss, but the mechanism is not straight-forward. Increased firing rates are observed from the cochlea, but these cannot be the direct cause. The sensation seems to be coincident not with the increased firing rates coming up to cortex, but with the establishment of synchronous firing in cortex.

A complex reorganization is indicated when a region of the cochlea experiences hearing loss. The tonotopic organization of auditory cortex is disturbed, which then allows an incoming signal to be misread as tinnitus. More specifically, reduced lateral inhibition in cortex gives rise to the excess synchrony in neuronal firing. The tinnitus may also be a kind of edge effect in that the tinnitus occurs at the frequency where the slope of the hearing loss with frequency is maximum. This likely follows from the fact that we observe an increase in spontaneous firing rate above and below the injured zone.

The whole cascade can only be explained through a complex interaction of upcoming and descending signals. The loss of return signal from cortex causes the receptive fields at the thalamus to enlarge. Under normal circumstances, the cortex can sharpen the receptive fields because of its large architecture of lateral inhibition. “The moment there is deafferentation, there is less lateral inhibition. There is down-regulation of inhibition…”
Excess neuronal synchrony follows, which is perceived as tinnitus.

Changes in the EEG power spectrum are also observed in cases of tinnitus. The idling rhythm of auditory cortex, the Tau rhythm (7-12 Hz) is diminished. Theta amplitudes are reduced, but over time a peak emerges in the 1-3 Hz region, along with an increase in synchronous oscillations. In normal individuals, the slow waveforms modulate the gamma activity in auditory cortex, and this modulation is lost in tinnitus clients. Tinnitus is also associated with state of arousal. And a strong comorbidity with PTSD is found that is not traceable to the original trauma or torture. Neurofeedback is able to modulate the regulatory loop that induces tinnitus. Clinical success is correlated with changes in both tau and delta bands. The alpha/delta ratio is also highly correlated with tinnitus sensation.

Auditory Evoked Potentials

George Zouridakis, Director of the Biomedical Imaging Laboratory at the University of Houston, presented on recent findings on analysis of evoked potentials. We are drawn to the ERPs because they provide more direct access to information related to brain function than are provided by baseline measures. A practical hindrance, however, has been that the signals need to be synchronously added together to yield adequate signal-to-noise ratio. This means the use of repetitive challenges over some period of time. Zouridakis showed that with the application of independent component analysis to the EEG one can extract useful information from single events. This makes ERPs potentially useful in a reinforcement paradigm, which has not been the case to date. It remains to be seen whether the necessary signal-to-noise enhancement can be accomplished in what we consider to be real time in the feedback world. The ICA procedure used here was an interative one, which does not readily lend itself to real-time applications.

The basic question in ERP research has actually not yet been answered satisfactorily. That is the issue of whether ERPs result from neuronal recruitment or from the rephasing of ongoing EEG activity under the duress of the stimulus. Zouridakis found that the ERP amplitudes did indeed depend substantially on the phase of the underlying EEG signal, and the time delay with respect to the stimulus showed such a dependence as well. That presents a problem in feedback. Perhaps we should just be content for the time being to be in a position to extract useful information from single-shot measurements. Even that may remain a challenge. The analysis software is proprietary.

Much more could be reported about the conference. But I will bring this newsletter to a close with a couple of non-technical observations. With the distressing decline of the dollar vis-à-vis the Euro, being ripped off by European taxi drivers was not as readily taken in stride as in prior years. There is an easy prosperity in Europe right now, and this state of well-being has had its fallout for the diet as well. The good life apparently means not having to eat your veggies, and over my time in Europe I felt increasingly vegetable-deprived. Eventually I just had to forage at a grocery store in Seville for some raws.

European life expectancies consistently exceed ours, which is surprising given the degree to which Europeans still smoke, consume alcohol, and relish an indefensible diet. There is much less obesity in Europe than here, however, the poor diet notwithstanding. Or perhaps another answer lies in the fact that Europeans use much less of what is turning out to be biggest killer of Americans: Medicine itself. (On the other hand, standard allopathic medicine just saved my life. Again. There is that.)

Siegfried Othmer, Ph.D.

Share your thoughts in the comments section below.

Leave a Reply