A Modular Activation/Coherence Approach to Evaluating Clinical/QEEG Correlations and for Guiding Neurofeedback Training: Modular Insufficiencies,Modular Excesses,Disconnections, and Hyperconnections

ABSTRACT

Current approaches to QEEG-guided neurofeedback involve efforts to normalize the abnormalities seen, without reference to the functional localization of the cortical areas involved. Recent advances in cortical neurophysiology indicate that specific brain areas are developed to perform certain functions (cortical modules). Complex brain functions require cooperation between modules, particularly during a learning situation. For example, the left prefrontal “activation module” must cooperate with one or both occipital “visual modules” to attend and see something on a chalkboard. To remember what has been seen, both temporal “memory modules” must cooperate with the visual modules for the image to be retained in short-term memory. If the connections between these modules are not functioning optimally, visual learning will be impaired. Decreased coherence (hypocoherence) indicates a decrease in functional connectivity between these modules, and increased coherence (hypercoherence) indicates an increase in functional connectivity between the modules. Neurofeedback can be used to normalize coherence between these modules, thereby improving the efficiency of their cooperation in the learning process. If coherence is less than normal, it is trained up. If coherence is more than normal, it is trained down. Three cases are presented where this approach has succeeded in remediating the client's symptoms.     

Exploratory Analysis

Abstract

QEEG variables (5 activation, 2 relationship variables, 19 locations and 5 bands up to 64 Hertz) were collected under three activation conditions (auditory attention, visual attention and listening to paragraphs) on 84 subjects, consisting of 32 mild head-injured subjects (no loss of consciousness) and 52 normals over the age of 14. Additional variables collected included years of education, time since accident, sex, handedness and Shipley Institute of Living measures of IQ, verbal and abstraction scores.

The results were encouraging for future development of a discriminant employing activation conditions, as the results varied from 88% to 100% correct classification. Very few of the variables, which distinguished the groups under the three conditions or were used in the discriminant analysis, were shown to increase as the time since accident increased. This result, tentatively, indicates little effect of time on the improvement in the electrophysiological functioning of the brain. Time (spontaneous improvement) does not appear to heal the brain in any significant manner.     

QEEG Correlates of Auditory-Visual Entrainment Treatment Efficacy of Refractory Depression

Introduction. It is well established that the number of people diagnosed and suffering from depression is on the increase. Many of these patients are not responsive to first-line pharmacological intervention or simply cannot use medications for other reasons. As such, there has been a growing need for nonmedication approaches to treatment. The purpose of this study was to examine the use of auditory-visual EEG entrainment (AVE) at a 14 Hz (beta) frequency to decrease symptoms of depression with corresponding changes in neurophysiology.

Method. Sixteen participants ranged in age from 20 to 67 years and were screened utilizing the Beck Depression Inventory–II (BDI–II) and broken into two groups of 8 (simulated, AVE treatment groups), with a cross-over design. Both groups were given the BDI–II and QEEG testing at baseline, 4 weeks following either AVE or simulated treatment, and then again after an additional 4 weeks and a switch in treatment in the cross-over design.

Results. Results revealed significant reduction of depression only after the 4 weeks on AVE therapy of the BDI–II scores (p > .01). QEEG scores adjusted for normal age deviations demonstrate significant EEG change scores over time in cortical regions associated with mood regulation.

Conclusion. The findings indicate that AVE therapy may be a viable nonmedication therapeutic intervention.     

Real-Time Changes in Connectivities During Neurofeedback

Introduction. Changes in quantitative EEG during and in response to neurofeedback (NF) training was explored in patients with traumatic brain injury (TBI). Data from 19 adults with a TBI of moderate mechanical nature, non-drug-related, and without severe posttraumatic stress disorder or seizure disorder were analyzed (14 male and 5 female).

Methods. EEG was evaluated before, during, and after ROSHI NF training. Data were collected as duplicate samples of 6 min each during eyes open and eyes closed conditions, but only the eyes closed condition was analyzed.

Results. Significant changes in connectivity occurred during and in response to NF training.

Conclusion. Results showed significant changes in real-time QEEG connectivity. An evaluation of a larger subject population will clarify gender differences in connectivity responses to NF training.     

Is It Better to Train Power First or Coherence First?

Introduction. This study was done to see to what extent power training would correct coherence abnormalities in head-injured patients and to what extent coherence training would correct power abnormalities in a similar group of head-injured patients.

Method. Ten patients had power training first, and 10 patients had coherence training first (4 protocols with 5 sessions/protocol in each case).

Results. Either power or coherence training first resulted in normalization of most power and coherence abnormalities. Coherence training first resulted in significantly more new power abnormalities (10/client vs. 5/client for new power abnormalities). Power training first resulted in significantly more new coherence abnormalities (6/client vs. 2/client).

Conclusion. We did not find a clear-cut advantage for doing either power or coherence training first. However, we would recommend a repeat QEEG after doing either power or coherence first, since most original abnormalities will have resolved and there are likely to be several new abnormalities to be remediated.     

Ant Colony Optimization Based Feature Selection Method for QEEG Data Classification

Objective

Many applications such as biomedical signals require selecting a subset of the input features in order to represent the whole set of features. A feature selection algorithm has recently been proposed as a new approach for feature subset selection.

Methods

Feature selection process using ant colony optimization (ACO) for 6 channel pre-treatment electroencephalogram (EEG) data from theta and delta frequency bands is combined with back propagation neural network (BPNN) classification method for 147 major depressive disorder (MDD) subjects.

Results

BPNN classified R subjects with 91.83% overall accuracy and 95.55% subjects detection sensitivity. Area under ROC curve (AUC) value after feature selection increased from 0.8531 to 0.911. The features selected by the optimization algorithm were Fp1, Fp2, F7, F8, F3 for theta frequency band and eliminated 7 features from 12 to 5 feature subset.

Conclusion

ACO feature selection algorithm improves the classification accuracy of BPNN. Using other feature selection algorithms or classifiers to compare the performance for each approach is important to underline the validity and versatility of the designed combination.     

Frontal EEG delta/alpha ratio and screening for post-stroke cognitive deficits: The power of four electrodes

This study analysed correlations between post-stroke, quantitative electroencephalographic (QEEG) indices, and cognition-specific, functional outcome measures. Results were compared between QEEG indices calculated from the standard 19 versus 4 frontal (or 4 posterior) electrodes to assess the feasibility and efficacy of employing a reduced electrode montage. Resting-state EEG was recorded at the bedside within 62–101 h after onset of symptoms of middle cerebral artery, ischaemic stroke (confirmed radiologically). Relative power for delta, theta, alpha and beta, delta/alpha ratio (DAR) and pairwise-derived brain symmetry index (pdBSI) were averaged; over all electrodes (global), over F3, F4, F7, F8 (frontal) and P3, P4, T5, T6 (posterior). The functional independence measure and functional assessment measure (FIM–FAM) was administered at mean 105 days post-stroke. Total (30 items) and cognition-specific (5 items) FIM–FAM scores were correlated with QEEG indices using Spearman's coefficient, with a Bonferroni correction. Twenty-five patients were recruited, 4 died within 3 months and 1 was lost to follow-up. Hence 20 cases (10 female; 9 left hemisphere; mean age 68 years, range 38–84) were analysed. Two QEEG indices demonstrated highly-significant correlations with cognitive outcomes: frontal DAR (ρ = − 0.664, p ≤ 0.001) and global, relative alpha power (ρ = 0.67, p ≤ 0.001). After correction there were no other significant correlations. Alpha activity – particularly frontally – may index post-stroke attentional capacity, which appears to be a key determinant of functional and cognitive outcomes. Likewise frontal delta pathophysiology influences such outcomes. Pending further studies, DAR from 4 frontal electrodes may inform early screening for post-MCA stroke cognitive deficits, and thereby, clinical decisions.     

Frontal EEG predictors of treatment outcome in major depressive disorder

ObjectiveTo investigate the role of frontal EEG as predictor of clinical response to SSRIs or venlafaxine in major depressive disorder (MDD).Method82 subjects (age 35.9 ± 13.0; 47.6% female) meeting DSM-IV criteria for MDD entered an 8-week prospective treatment with SSRIs or venlafaxine. At baseline and week 1 we recorded serial, 4-channel EEGs (F7-Fpz, F8-Fpz, A1-Fpz, A2-Fpz). We evaluated prospectively the relative theta power as predictor of treatment outcome. We also developed an Antidepressant Treatment Response (ATR) index using EEG parameters assessed at baseline and week 1.Results45 subjects (54.9%) responded to treatment (HAM-D-17 reduction ≥ 50%). At baseline, frontal relative theta power (i.e., 4–8 Hz power/2–20 Hz power) was significantly (p = 0.017) lower (21%) in treatment responders than in non-responders (24%). Baseline relative theta power predicted treatment response with 63% accuracy [64% sensitivity, 62% specificity, 66% area under the receiver operator curve (AUROC) (p = 0.014)]. Relative theta power at week 1 predicted treatment response with 60% accuracy [62% sensitivity, 57% specificity, 61% AUROC (p = 0.089)]. ATR predicted response with 70% accuracy [82% sensitivity, 54% specificity, 72% AUROC (p = 0.001)].ConclusionUsing automated analysis of frontal EEG collected during the first week of antidepressant treatment it may be possible to facilitate prediction of SSRI or venlafaxine efficacy in MDD.     

Quantitative topographic electrophysiology and functional neurologic status in right middle cerebral artery infarction

Thirteen patients with clinically and radiographically defined right middle cerebral artery infarction were studied using EEG, quantitative electroencephalographic (QEEG) spectra, and multi-channel evoked potentials. The purpose of this effort was to develop QEEG rules that related to the patient's neurologic status. Three QEEG relative delta spectral patterns were identified in the right hemisphere which related to neurologic residua. These include limited perisylvian involvement, mixed involvement of perisylvian and extrasylvian regions, and extrasylvian involvement only. While there were parallels between QEEG spectral patterns and auditory, visual and somatosensory evoked potentials, there were modality specific features consistent with functional differences.     

Poststroke QEEG informs early prognostication of cognitive impairment

Cognitive impairment is a common consequence of stroke, but remains difficult to predict. We investigate the ability of early QEEG assessment to inform such prediction, using binary logistic regression. Thirty‐five patients (12 female, ages 18–87) suffering middle cerebral artery, ischemic stroke were studied. Resting‐state EEG was recorded 48–239 h after symptom onset. Relative power for delta, theta, alpha, and beta bands, delta:alpha ratio, and peak alpha frequency were analyzed. Montreal Cognitive Assessment (MoCA) was administered, where possible, on day of EEG and at median 99 days (range 69–138) poststroke. Eight patients could not complete the baseline MoCA, and four the follow‐up MoCA, for varying reasons (most commonly, stroke symptoms). Fifteen patients (48%) had cognitive impairment (MoCA score ≤25) at follow‐up. One QEEG index was able to correctly predict presence/absence of cognitive impairment in 24/31 patients (77.4%), whereas predischarge MoCA did so in 23 patients. This index, relative theta frequency (4–7.5 Hz) power, was computed from only three posterior electrodes over the stroke‐affected hemisphere. Its predictive accuracy (three electrodes) was higher than that of any “global” QEEG measure (averaged over 19 electrodes). These results may signify association between poststroke alpha slowing and cognitive impairment, which may be mediated by attentional (dys)function, which warrants further investigation. Pending further studies, QEEG measure(s)—from a few electrodes—could inform early prognostication of poststroke cognitive outcomes (and clinical decisions), particularly when cognitive function cannot be adequately assessed (due to symptoms, language, or other issues) or when assessment is equivocal.