LabVIEW Automates Brain Wave Experiments in the Neurophysiology Lab

Introduction
Brain potentials are divided into sponta-neous and event-related. The spontaneous result from the regular brain activity (EEG potentials). The event-related potentials (ERP) result from external brain excitation (events) and are divided into evoked and anticipatory. Evoked potentials appear after excitation as a reflex of the brain. Anticipatory potentials appear before the corresponding event and represent an expectation of the same and usually a motor preparation process for it in the brain. The most prominent example of the expectation-related potential is the contingent negative variation (CNV) potential. It is extracted from the subject’s EEG within the CNV experiment.

The CNV experiment is based on the CNV paradigm, which applies two brain stimuli (S ₁ /S ₂ ) to the subject with a constant interstimulus interval. S ₁ is a warning stimulus, and S ₂ is an imperative stimulus to which the subject must react. The procedure is repeated tens of times, during which an ERP, produced in the EEG trace between the stimuli, shapes itself toward a specific CNV wave. The ERP after 10 to 20 trials can clearly show both components - the evoked (short) potential due to S ₁ as well as the anticipatory (late, expectancy) potential together with the preparatory potential prior and due to S ₂.

The dynamic CNV (DCNV) experiment is an extension of the CNV experiment. The extension involves switching S ₂ on and off, which occurs automatically after fulfilling certain conditions in the experiment’s environment, thus forcing a cyclic process of building and degrading of the CNV wave. The subject is not informed about the nature of both stimuli, so the expectation of appearance (absence) of S ₂ during the experiment completely corresponds to the learning process.

The CNV wave (extracted ERP) can be qualified by one of its parameters, such as amplitude or slope. After the experiment, a statistical curve of the qualifying parameter is drawn across the trials. This statistical curve is denoted as the electroexpectogram (EXG) and directly presents the subject’s cognitive capabilities.

Experiment Design
"AEP Research Tool," written in LabVIEW, embraces the acquisition, signal processing, and analysis of the EEG and EOG traces (latter used for validation of the EEG against artifacts) as well as reporting. The hardware used in the experiment consisted of a bandpass amplifier for mV ranges, an AT-MIO-E board, a sound card for applying the stimuli, and a push button switch with a TTL interface (figure 1).
The system acquires two differential analogue channels (EEG and EOG). We use audio with S ₁ being a short 1 kHz warning beep and S ₂ being a longer 2 kHz imperative beep. It is essential that the subject is aware neither of the nature nor of the number of the stimuli. The acquisition lasts for 7 s. S ₁ is issued in 1second into the acquisition; S ₂ is issued in 3 s if applied by the algorithm. During the experiment, the subject learns about the number, nature, and order of the stimuli, thus demonstrating the process of learning by shaping the ERP wave toward the expected CNV. The subject has to react upon hearing S ₂ by pressing the button and immediately stopping it. This is to prevent falling asleep and lowering of concentration. The number of trials in the experiment is set to maximum 100 successful (120 trials total). The gap between two consecutive trials varies from 12 to 15 s to avoid timing determinism. The criterion for ERP being a CNV is defined above. After three consequent CNVs are detected, S ₂ is turned off, and the subject learns to forget the imperative stimulus, thus lowering the value of the CNV-qualifying parameter. After three consequent not-CNVs, S ₂ is turned on again, and so on. The EOG trace is used for automatic validation of the EEG trace against artifacts defined as voltage sequences longer and higher than preset thresholds. There is a second manual criterion applied, where the operator can reject current EEG if artifacts are recognized visually for 3 s after the acquisition. Rejection of such trials is necessary since the process of extraction of the ERP uses a cumulative iterative FIR filter that averages the acquired signal by ansamble, so every artifact that passes it will influence the extracted ERP till the end of the experiment.

Algorithm Design
The software package written in LabVIEW works under Windows NT/95. Main panel (figure 2) shows the acquired EEG signal (current trial), the extracted CNV potential and its linearized model, as well as the required measurements and calculated values. A vertical marker on the CNV Morphology represents the reaction time. All of the subject’s acquired and calculated data is saved in an ASCII file. The optimal filter is defined as follows - CNV i = d = CNV i - 1 + c • EEG i "AEP Research Tool" is completely hardware-synchronized. Acquisition is onboard clocked. The audio stimulation is performed through a sound card. The reaction time is measured by an onboard counter, started by a digital output from the card issuing pulse at the same moment with the start of S ₂ and stopped by the user pressing the button or the time-out pulse, applied again by the same digital line.

Conclusion
The EXG curve represents a cognitive wave obtained from the human brain showing the oscillatory change of the expectancy status in the human brain and is a manifestation of the expectation process and the learning process taking place in the human brain, during the DCNV paradigm.

"AEP Research Tool" successfully performs the expected task and produces an excellent nine page report containing all required statistics. Experiments during the test phase indicated that different categories of subjects (healthy adults, groups of people with distinctive neurological disturbances, and little children) may produce quite different electroexpectograms, but very similar within the groups, because some are not able to raise their CNV waves above the thresholds. The design of this experiment illustrates unlimited possibilities of LabVIEW-based systems in the medical field.

For more information, contact:

Roman Golubovski

Electrical and Biomedical Engineer

116-3/24, MK-1000 Skopje, Macedonia

Email: roman.golubovski@iname.com

Tel: + 389 (0)91 165 367

Voice/fax: + 389 (0)91 165 304