**3. Results**

The two parameters selectable by the clinician are:

**a.** position of the electrodes, according to symptom profile and symptom changes;

Until 2006, the signal processing was very similar to the classical beta-SMR scheme [44]. However, the reward frequency setting of a 3 Hz wide variable bandpass-filter was useradjustable over the entire conventional EEG spectrum from 1.5 to 40 Hz in center frequency. For that purpose, a horizontal slider was implemented in the graphical user interface. The inhibits were comprised 10 separate filter blocks in fixed frequency steps in the range between 1 and 40 Hz. For both the reward and inhibit scheme, threshold setting was auto-corrected to

Specific design parameters in the signal-processing chain between initial EEG acquisition and ultimate feedback animation have always been assessed and optimized by means of an empirical approach based on qualitative evidence criteria. (A useful analogy to this process is the optimization of the suspension system of a car, where human factors come prominently

By expanding the underlying model of neurofeedback to incorporate the current understanding of the brain as a self-organizing dynamical system that interacts with itself by means of neurofeedback, improved approaches to signal processing and coupling to the feedback animations have been sought. This process got underway in 2001. In that regard, also slow cortical potentials were investigated. With the availability of greater computer power for additional signal processing as well as advanced signal acquisition technologies, it was found that the addition of such slow potentials appears to offer the brain a more direct and effective feedback interaction. It turned out that also with this scheme, tailoring of the parameter setting to the individual patient is beneficial or even necessary, just as was previously found for frequency-band training in the conventional EEG spec-

In contrast to the classical concept of a rewarding experience that is controlled by the amplitude of the EEG in a given frequency band, the goal here is to present the brain with the most relevant representation of its slow cortical potential. For that purpose, derivations from the measured signal control various features in the feedback animation in a way that optimizes

For the purpose of continuity of the clinician's experience with the earlier era, the terminology of "reward frequency" was retained, as the rules for settings and for the optimization procedure carried over into the ILF region. However, the absence of discrete rewards in the ILF training meant that the traditional terminology of reward had lost its meaning. The unfolding of the continuous ILF signal allowed for no external reinforcers. Additionally, the slider that controlled the target frequency within the EEG regime was retained in the new design, but its function in the ILF regime must be understood differently. With the adopted

**b.** adjustment of the reward frequency according to patient feedback.

maintain a chosen level of difficulty, the "percent success."

into play.)

80 Biofeedback

trum [45].

the brain's opportunity to engage with them.

After completion of 20 NFB sessions, all participants indicated improvement of their state. Most of them noticed a decrease of inner tension and reactivity to stressful factors. Further, they reported on stability of mood, improved body and space awareness, increase of energy level and of cognitive performance.

The post-training EEG patterns in all eight subjects revealed significant enhancement of spectral power in 0–0.5 Hz frequency band compared to the pre-training EEG. The locations of the most prominent changes were different: in some subjects, the dramatic ILF power increase was observed over the frontal-central region, in other cases over the posterior brain areas.

**Figure 1** presents the EEG recordings before and after ILF NF course in one of the participants of our study.

**Figure 2** demonstrates the increase of the level of infra-slow activity in 0.03–0.05 Hz range in the post-training EEG in this participant. This increase is most prominent over frontal region.

Two-way ANOVA revealed a significant main effect of the factor "Condition" for the slow activity in 0–0.5 Hz frequency band F [1,7] = 18.4, p < 0.01. This effect is illustrated in **Figure 3**.

**Figure 1.** Pre-training (on the left) and post-training (on the right) EEG in the 43-year-old male subject. EEG recorded in the linked ear montage/reference during VCPT performance. Scale: 200uV/cm, speed–1.875 mm/s, time constant–10.0 s (0.016 Hz), low frequency filter–0.5 Hz.

An increase of the logarithmical power averaged in eight subjects in 0–0.5 Hz band is seen in

**Figure 3.** Influence of the ILF NF on the power of EEG activity in 0–0.5 Hz frequency band. X-axis**–**Electrodes localizations;

Effect of Infra-Low Frequency Neurofeedback on Infra-Slow EEG Fluctuations

http://dx.doi.org/10.5772/intechopen.77154

83

Y-axis–Logarithmic scale of ILF (0–0.5 Hz) power. Whiskers represent a 95% confidence interval.

All participants had normal mental and physical development and had no history of any neurological abnormalities. However, most of them reported some form of self-perceived psychological and physiological issues such as fatigue, depressed mood, symptoms of inner tension, mood swings, headache, and sleep problems. These were accompanied by cognitive concerns

After 20 sessions of ILF training, the pattern of ILF activity at rest changed dramatically. The main difference was an increase in the amplitude of the ILF activity up to 0.3–1.0 mV in all recording sites. These results indicate that ILF training modified the baseline brain state in each case. It is important to add here that the changes in brain dynamics were associated with improvement

all 19 electrodes localizations.

such as diminished attention or poor working memory.

**4. Discussion**

**Figure 2.** EEG power spectra at Fz in the 43-year-old male subject before and after training**.** Pre-training–Lower curve, post-training–Upper curve, X-axis–Frequency and Y-axis–Spectral power in logarithmic scale.

Effect of Infra-Low Frequency Neurofeedback on Infra-Slow EEG Fluctuations http://dx.doi.org/10.5772/intechopen.77154 83

**Figure 3.** Influence of the ILF NF on the power of EEG activity in 0–0.5 Hz frequency band. X-axis**–**Electrodes localizations; Y-axis–Logarithmic scale of ILF (0–0.5 Hz) power. Whiskers represent a 95% confidence interval.

An increase of the logarithmical power averaged in eight subjects in 0–0.5 Hz band is seen in all 19 electrodes localizations.
