General Methodology Involved In Neurotherapy Research
All findings protected by Patent # 6309361 B1 Method for Improving Memory by Identifying and Using qEEG Parameters Correlated to Specific Cognitive Functioning - issued 10-30-2001
METHOD
Theoretical Rationale of the Research
This Web site presents information on the result of EEG biofeedback interventions in the improvements of IQ scores on standardized testing and memory improvement scores. The ability of the brain to respond to operant conditioning techniques (long established in the psychological literature) is established in these studies and others. Operant conditioning involves providing a reward after the behavior has been demonstrated. The organism, once rewarded, will engage in that behavior again. The ability of the brain to respond to this conditioning paradigm on the level of specific electrophysiological variables opens up a potential goldmine of treatment options. The goal of this research was to ascertain if there were specific patterns which could account for success on specific cognitive tasks. For example, does our ability to recall what someone said to us depend on particular type of qEEG activations and connections. If it can be shown that there are particular patterns for particular tasks, then it would theoretically be possible to set up operant conditioning protocols to improve those variables under those task conditions and thus improve a subject's ability in that area. The ability of the brain to be malleable and responsive allows us to potentially change the functioning of the brain. This goal has been realized in the data presented on auditory memory. The parameters which are responsible for auditory memory were discovered in this research and employed in the treatment of the subjects presented, with dramatic results. Many of the discoveries resulting from this research have been published, accepted for publication or submitted for publication. Due to the lengthy time between completed research and publication, the researcher has decided to place the results of some of the research on this web site.
Subjects
A total of 170 subjects underwent cognitive activation tasks while wearing an electro-cap which allowed the recording of all of the Quantitative EEG (qEEG) variables to be recorded. The tasks involved an eyes closed condition, auditory and visual attention, auditory memory for word lists and paragraphs, memory for names of faces, reading material, problem solving, visualization, mathematical problems (multiplication tables and two digit spatial addition), silent pronunciation of nonsense words, hearing words, spelling, recall autobiographical information and experiencing the emotions of love, happiness and sadness. All the memory tasks involved both a short and delayed recall period (30-45 minutes). The group was composed of individuals who had experienced a brain injury, normals (with no history of neurological problems, brain injury or learning disability) and individuals with a history of learning disability. The qEEG variables were correlated with performance on the tasks in terms of the absolute level of the variable and the degree of activation from a relevant condition. For example, the hearing of paragraphs involves both auditory attention and auditory memory ability. To separate out the auditory attention aspects of listening to paragraphs, the qEEG values of a subject's auditory attention response (hearing taps on a table) was subtracted from hearing paragraphs. In this manner, the subject's activation of a qEEG variable above the auditory attention condition could be obtained and correlated with their memory performance. The results published on this Web site employ only the normal subjects, except where indicated. The adult sample consists of individuals over the age of 13 and the child samples involved subjects under the age of 13. The sample size obtained with each task in the research is indicated on the figures, as well as whether it involves the absolute level or the degree of activation variables. For both the adults and children, only the right handed subjects were selected.
Apparatus
While the subjects underwent the experiment, they were videotaped and recorded. The recording device combines a computer, video, audio recording equipment and allows an experimental recording session to be saved to a hi 8 mm tape with all pertinent information. The videotape, which is saved, is a split screen videotape, with the left side reflecting the EEG recording with the appropriate epoch number and the right side of the screen showing the subject during the experiment. The epoch number refers to a one-second period of time. Thus, epoch number one is the first one-second period of the recording and epoch number sixty is the sixtieth second of the recording. This device enables the experimenter to review the tape to check and confirm the scoring of the subject's responses during the experiment. The EEG recording equipment of Lexicor Medical Technology, Inc. (NRS-24) was employed. The sampling rate was set to 256 to allow for examination of up to the 64-Hertz range with a 60-Hertz notch filter. In the system employed in this study, filtering is accomplished in the software. The signals passed are between .5 and 64 Hz (3dB points). The signals, which pass, are then subjected to a Fast Fourier Transform (FT) using Cosine-tapered windows which output spectral magnitude in microvolts as a function of frequency. The bandwidths were divided according to the following division: Delta: .5-3.5 Hertz, Theta: 4-7.5 Hertz, Alpha: 8-12.5 Hertz, Beta1: 13-31.5 Hertz, Beta2: 32-63.5 Hertz. This equipment provides for the collection of data in the standard 10-20 system (ear linked references) format of EEG data collection. Impedances below 5 Kohms were obtained on all locations. Gain was set to 32000 and the high pass filter was set to off. The earlobes and forehead were prepped with rubbing alcohol and Nu-Prep. An Electro-cap was employed and spaces filled with Electro-gel. The data was visually analyzed and marked for deletion when artifact was evident. All measurements available through the software provided by Lexicor Medical Technology, Inc. were employed. These included the following for each bandwidth and employed the peak-to-peak method.
Activation Measures
Absolute Magnitude: the average absolute magnitude (as defined in microvolts) of a band over the entire epoch (one second)
Relative Magnitude: the relative magnitude of a band (absolute magnitude of the particular band divided by the total microvolt generated at a particular location by all bands)
Peak Amplitude: the peak amplitude of a band during an epoch (defined in microvolts)
Peak Frequency: the peak frequency of a band during an epoch (defined in frequency) Symmetry: the peak amplitude symmetry between two locations in a particular bandwidth- i.e. defined as (A-B)/(A+B).
Connectivity Measures
Coherence: the average similarity between the waveforms of a particular band in two locations over the one-second period of time. Conceptualized as the strength/number of connections between two positions.
Phase: the time lag between two locations of a particular band as defined by how soon after the beginning of an epoch a particular waveform at location #1 is matched in location #2.
PROCEDURES
For each of the conditions a description is provided of the task (in the accompanying results section). Data Analysis. The raw data, once artifacted, was analyzed with the Exporter program available from Lexicor Medical Technology. The Exporter software solves the cumbersome time problem of obtaining the required figures from the raw data file. The program generates the values for the variables under consideration from the raw data file and generates ASCII, comma delimited files, which can then be imported into Excel or CSS Statistica. One CSS spreadsheet was generated for each subject. Every epoch of the spreadsheet was labeled according to the experimental condition, which was occurring at the time. Once the labeling was completed, the file underwent a series of t-test comparisons, i.e. auditory attention versus listening to paragraphs. A t-value was generated for all 2,945 variables for each of the comparisons conducted. These comparisons were placed into Excel spreadsheets. Each subject had an Excel spreadsheet containing all of the comparison results. Once all the comparisons were completed, the results from the Excel spreadsheet were transferred to a CSS statistica file containing similar comparisons for the other subjects. Thus, one CSS file would contain the all of the information for listening to paragraphs. The CSS files for each comparison would consist of three files. For example, in the immediate recall of paragraphs, one file would contain the absolute levels of the condition, another file would contain the degree of activation values from the relevant comparison condition (eyes closed), and a third the t-values of the differences between eyes closed and silent recall (not presented in results). Some of the figures indicate fifty-eight or fifty-nine subjects, as the data was not available for all conditions for all subjects.
STATISTICAL CONSIDERATIONS
The Bonferroni correction was not considered appropriate for analysis as it fails to take into account different categories of information. Setting the Alpha level to .05 would result in 147 significant findings (of the 2945 variables under consideration) by chance alone. To reduce this statistical problem to manageable levels the following were considered.
Epochs with minor Delta activity (defined as between 40-60 microvolts) were included in the data analysis for two reasons: (1) Eye movement (the predominant Delta activity artifactual concern) may relate to cognitive functioning as rightward eye movements have been shown in some studies to activate left posterior cortical areas and vice-versa; and (2) the need to obtain as many acceptable epochs in the short 30 second period of time as possible to increase statistical power. However, actual Delta band activity from these epochs was not included in the analysis for two reasons: (1) the inclusion of Delta activity could be misleading as to the nature of the underlying brain activity; and (2) reading-related effective qEEG parameters were focused in the higher bandwidths, thus minimizing the probability that Delta activity would be found to be of importance in this study. The effect of this decision will be discussed in the results section.
Statistical significance was considered in view of the parameter under consideration. For example, with one of the activation measures (relative power, etc.) there are 19 locations and 4 bands (excluding Delta), resulting in 76 possible significant findings. An Alpha level of .05 would produce approximately 3.8 significant findings by chance alone. Significant activations are reported if there are 1–two or more in adjacent positions, 2– involve different bands in the same location or 3–are relevant to the phase and coherence activity. The total number of activation variables resulting from 19 locations, 4 bandwidths and 4 parameters (excluding symmetry) is 304. The symmetry measures produce 674 variables. The total number of connection measures resulting from 19 locations, 4 bandwidths and two parameters is 1368. The resulting total number of variables (after removal of Delta related variables) under consideration is 2346. The number of symmetry measures was reduced by summation of a location's symmetry relationship to the 18 other locations (for the Beta1 and Beta2 bands only). Thus each location would have two symmetry measures for the Beta bands activity.
The connection measures (phase and coherence) were reduced by examination of the four frequencies emanating from each location to all other locations. Conceptually this is equivalent to considering that a location is capable of "generator capacity" in the different frequencies. The concept of generators has been employed previously in EEG research (Moran, Tepley, Jacobson & Barkley, 1993). This pattern of analysis assumes a cortical generator for the separate bands. For example, the phase Alpha value from F7 to all 18 other locations was summed. This method of data reduction was conducted for all four frequencies (Theta to Beta2), both connection variables (phase, coherence) and all locations. Thus, each location would have eight possible significant relationships (phase Alpha, coherence Alpha, etc.). Any significant relationship from a location (alpha set at .05) would be above chance levels.
Description of Figures
There are two sets of figures for each of the conditions (study, immediate and delayed recall). The first set of figures for each of the comparisons presents the absolute level of activation for the variables, which are related to success. The second set of figures for each of the comparison presents the degree of activation from the subject's relevant comparison condition (visual attention or eyes closed).
The black filled circles represent the locations activated (magnitude, relative power, peak frequency, and amplitude asymmetry) to a significant level according to the previous discussion of significance. The lines represent the significant phase and coherence values between locations. A black filled circle represents the "origin" of these lines. For ease of visual inspection, some of the circles are not drawn where the line terminates. The effective (positive relationships) and ineffective (negative relationships) parameters are noted on each set of figures. The label on top of each head figure indicates the variable that the head figure is referring to, according to the following nomenclature. Each figure is accompanied by text that presents the relevant variables.
T-Theta: A-Alpha: B1-Beta1: B2-Beta2: M-Absolute Magnitude: R-Relative Magnitude:
PKA-Peak Amplitude: PKF-Peak Frequency: Sym-Symmetry: P-Phase: C-Coherence
The following examples are provided for clarification.
PA - reflects phase Alpha: CT – reflects coherence Theta: RPB2 – reflects relative power of Beta2: MB1 – reflects magnitude of Beta1: PKAA – reflects peak amplitude of Alpha: PFKB2- reflects peak frequency of Beta2: SymB1- would reflect symmetry of Beta1 band (Symmetry measure employ a combination method, where a particular location’s symmetry measure is calculated in reference to all other positions and is calculated only for the beta bands. This method of calculating the symmetry method was to aid in reducing the number of variables under consideration. Only Beta activity was considered in these calculations, as it is generally considered relevant to task conditions.)
RESULTS
Review of Research
Despite neuroscience being a new and rapidly developing scientific area, there is very little literature on the questions being raised in this research.