Biomedical Instrumentation Lab
Visual Evoked Response (VER)
1. Tasks
The purpose of this laboratory is to:
- Introduce you to the EEG, and the VER in particular
- Measure and record data for a subject’s EEG (alpha wave in particular)
- Measure and record data for a subject’s VER, when presented with a visual stimulus (strobe light)
- Process the EEG and VER data using MATLAB
You are expected to complete the laboratory according to the following guidelines. All experimental work including raw and processed data and descriptions should be written in a bound laboratory book (NOT LOOSE PAGES!).
There are tasks that are also expected to be completed – usually after the experimental acquisition part of the laboratory. In many cases these tasks can be done after the formal lab session if you are running short on time. In most cases you will need to analyse some previously collected data using MA TLAB.
You are to write a formal lab report (3-4 pages) The practical report should explain what were the objectives of the experiments. It should describe the methods and equipment employed; reference to the instruction sheets is adequate here, but all individual departures from the procedures or equipment envisaged in the instructions should be noted in detail. Include any negative or incomplete results of experiments. It should refer to pertinent information sources such as literature and instruction manuals. It should note any insights derived from the analysis process. At the end of the data analysis it should draw conclusions.
2. Theory
2.1 Visual Evoked Response (VER)
Electroencephalograms (EEGs) are traces of measured bio-potentials from electric activity in the brain. Somatosensory evoked potentials are bio-potentials created in response to neural activity caused by sensations and perceptions. These bio-potentials can be recorded as specific types of EEGs, such as the auditory evoked response/potentials (AER or AEP) and visual evoked response/potentials (VER or VEP).
In general, with scalp electrodes applied to a relaxed adult, electrical activity can be observed with a dominant frequency of approximately 8-13 Hz and amplitude of approximately 20-200 μV. This activity is called the alpha wave. It is most prominent in the occipital lobe (primary visual cortex), when the subject’s eyes are closed (e.g. sleeping) and the subject is relaxing (e.g. not concentrating). Other waves measured with an EEG are the beta, theta, and delta waves. Note that the EEG in general is greatly affected by the cerebral state of the subject.
EEG recordings in general, and VER recordings in particular, are difficult to obtain due to the small signal magnitudes in the presence of large noise signals. A technique called ensemble averaging is used to obtain an average recording; this technique assumes that the evoked response (e.g. VER) is synchronised with the stimulus, whereas the background EEG activity is a random signal.
The VER typically requires electrode recordings from the occiput (back of the head), with a complex pattern of waves observed from 0.1 ms to several 100 ms after the stimulus that caused the evoked response. The clinical use of the VER is to test the integrity of the afferent neural circuit associated with the optic neural pathway. The VER is often used with an electroretinogram (ERG), a biopotential recording made with contact lenses, used to assess retinal damage (as opposed to damage in other areas of the optic neural pathway).
2.2 Further Reading
Webster, J. G. (Ed)., Medical Instrumentation: Application and Design, 3rd Ed., USA, John Wiley & Sons, pp 151-175.
Germann, W. J. & Stanfield, C. L., Principles of Human Physiology, USA, Benjamin Cummings, pp. 251-254.
3. Equipment
- 3.1 3 cup electrodes
- 3.2 IEC xenon strobe light
- 3.3 Drape
- 3.4 PowerLab 26T data acquisition device and bioamp
- 3.5 LabChart acquisition software
- 3.6 Alcohol swabs
- 3.7 Conductive paste
4. Procedures
4.1 Electrodes
Place three electrodes on the test subject in the following locations:
- Low on the forehead (positive lead)
- Back of the head at the top of the occipital bone (negative lead):
- On the bony part behind one of your ears (isolated ground electrode)
The electrodes can be placed directly on the skin, however, to increase the signal quality, preparation of the skin prior to electrode adhesion is beneficial. Alcohol cleansing swabs are available to improve the contact and reducing the recording artifact. Conductive paste is also used to bridge the gap between the electrode cup and the skin.
4.2 Electrode Setup
Attach the positive and negative electrodes to the CH1 inputs on the yoke of the PowerLab electrode lead cable. Connect the reference electrode ground electrode to the EARTH/COMMON input. Connect the other end of the cable to the input to the bioamp.
4.3 PowerLab / LabChart Setup
Open LabChart and go to “Setup” → “Channel Settings…”. Turn off everything except Channel 1 and Channel 3. Set the input range on Channel 3 to 200 μV. Ensure the input range to Channel 1 is 10 V. Set the sampling rate on both of these channels to 10 kHz.
Analog Output 1 on the PowerLab device should be connected to Analog Input 1 and also (using a splitter) to the external trigger input of the strobe light; the analog output will be used to make the strobe light flash and the analog input will sample this trigger signal, so the VER responses can be synchronised later when an ensemble average is taken. All of these connections should already be done on the workbench.
Go to “Setup” → “Stimulator…” and set up a repeating pulse pattern which sends a trigger pulse to the strobe light every 2 seconds, repeats 50 times (in practice this would be done hundreds of times), with a pulse width of 0.1 seconds and a height of 5 V. Ensure the “On” button is pressed and that is it set to start when sampling starts.
Start a test recording. By observing the waveform, you may wish to add a notch filter to Channel 3, and also perhaps use Channel 5 to create a band-pass filtered version of the result, with cut-off frequencies between approximately 1 Hz and 300 Hz.
4.4 Alpha Wave Measurement – Without Noise
Temporarily disconnect the strobe light trigger input. Effort must be taken to minimize the noise from the recordings, as the combined gain of the amplifier system is approximately 10,000. To do so, ensure that the power cables are as far as possible from the electrode and amplifier cables, and that the electrode contacts are good.
The subject should close their eyes and relax. An alpha wave should be observed on Channel 3 in LabChart. When you are ready you should start a new LabChart document and save this recording here. Record approximately 10 seconds worth of data. You can also export the data in MATLAB format for analysis later.
Store the data using a filename that makes sense in one of your home directories, so that it will be accessible after the lab is complete. You will need this file (and all other data files collected in this lab) for data processing with MATLAB.
4.5 Alpha Wave Measurement – With Noise
Disconnect the isolated ground electrode lead and repeat the same measurement. This will introduce significant noise to the measurements. Be sure to start a new document with a different name for this experiment.
4.6 Muscle Movement Effect on the EEG
Connect the isolated ground electrode, to reduce the signal noise. Be sure again to start a new document with a different name (so you do not loose the previous data). You do not need to retain the data for this step, so a temporary file name is fine.
Re-run the measurement program, but this time the subject should blink during the measurement (e.g. 1 s into the measurement). Do you notice a difference on the measurement? Try introducing an auditory stimulus (e.g. snap a finger near the subject’s ear).
4.7 Measure the VER
Re-connect the external trigger to the strobe light. Place the drapes over the subject’s head and allow the subject to adapt to the low light (e.g. a few minutes). Record the data until the strobe light stops flashing. Again, save to a new data file for use later. The subject should try not to blink and definitely not within 0.5 seconds either side of the stimulus.
5. Discussion
Using MATLAB, process the EEG and VER data as follows:
5.1 Alpha Wave With and Without Noise
Perform a power spectrum analysis on the data, to determine the frequency components of the wave. The MA TLAB pwelch command can be used for this task (if the frequency resolution is too poor then increase the window size). Plot the power spectrum. The signal and noise components of the recorded wave should be present. The function prototype is:
[pxx,w] = pwelch(x,window,noverlap,nfft)
x is the EEG signal being analysed. Try the following settings for the rest (assuming a sample rate of 10 kHz):
window = hamming(10000);noverlap=5000; nfft=2^18;
5.2 VER
Write a MATLAB program to segment the exported EEG data from PowerLab; create a matrix with each row containing the response to one flash of the strobe light (should be 50 flashes) and the columns containing about 0.5 seconds of data before the light strobes and 1 second of data which followed the strobe (1.5 s of data per flash). The stimulus onset of each run is not necessarily the same, therefore once the data is aligned in this way, the measured results from all runs can be averaged (ensemble averaging across the 50 rows). In real experiments this may also involve viewing each run and removing those that have excessive artifact e.g. because of the subject blinking.
Using MATLAB, plot the averaged data on a graph with at least half a second of pre-stimulus information, and the time axis scaled in seconds with t=0 representing the onset of stimulus. Graphically compare the average response with the first stimulus run, to see the effect of signal averaging.
Try to identify any evoked response waves. Is the response always causal? If not, why might it not be? You should also examine the effect of low-pass filtering the data before averaging.
In an electrically noisy laboratory environment these measurements are difficult to make. We may provide you at a later date with a sample data file showing a good response.