Source: Laboratories of Jonas T. Kaplan and Sarah I. Gimbel—University of Southern California
Given the overwhelming amount of information captured by the sensory organs, it is crucial that the brain is able to prioritize the processing of certain stimuli, to spend less effort on what might not be currently important and to attend to what is. One heuristic the brain uses is to ignore stimuli that are frequent or constant in favor of stimuli that are unexpected or unique. Therefore, rare events tend to be more salient and capture our attention. Furthermore, stimuli that are relevant to our current behavioral goals are prioritized over those that are irrelevant.
The neurophysiological correlates of attention have been experimentally examined through the use of the oddball paradigm. Originally introduced in 1975, the oddball task presents the participant with a sequence of repetitive audio or visual stimuli, infrequently interrupted by an unexpected stimulus.1 This interruption by a target stimulus has been shown to elicit specific electrical events that are recordable at the scalp known as event-related potentials (ERPs). An ERP is the measured brain response resulting from a specific sensory, cognitive, or motor event. ERPs are measured using electroencephalography (EEG), a noninvasive means of evaluating brain function in patients with disease and normally functioning individuals. A specific ERP component found across the parietal region of the scalp, known as the P300, is enhanced in response to oddball events. The P300 is a positive-going deflection in the EEG signal that occurs about between 250 and 500 ms after stimulus onset. In general, early potentials reflect sensory-motor processing while later potentials like the P300 reflect cognitive processing.
In this video, we show how to administer the oddball task using EEG. The video will cover the setup and administration of EEG, and analysis of ERPs related to both control and target stimuli in the oddball task. In this task, participants are set up with the EEG electrodes, then brain activity is recorded while they view control stimuli, interspersed with target stimuli. The procedure is similar to that of Habibi et al.2 Each time a target stimulus is presented, the participant presses a button. When the ERPs are averaged across the control and target stimuli, the neural correlates of each event can be compared in a selected time window.
1. Participant recruitment
- Recruit 20 participants for the experiment.
- Make sure that the participants have been fully informed of the research procedures and have signed all the appropriate consent forms.
2. Data collection
- EEG preparation (Note: These steps are for use with the Neuroscan 4.3 system with Synamps 2 amplifier and a 64-channel quick cap.)
- Participants in an EEG study should not have any hair products (e.g., gel, mouse, or leave-in conditioner) in their hair prior to their participation.
- Fill 2-4, 10-ml syringes with conductive electrode-gel (i.e., Quick-gel). It is suggested to stir the gel before using it to release air bubbles.
- Brush hair and scalp thoroughly (about 5 min).
- Clean head with alcohol and cotton gauze. Also clean the skin for placement of electrodes: two mastoids (behind each ear), below and above the left eye VEO (vertical electro-ocular), and the far sides of each eye HEO (horizontal electro-ocular; Figure 1, left).
- Using two-sided adhesive disks, place the electrodes.
- Measure the head from the front (directly between the eyebrows, mid-eye) to the inion (below the bump of the head in the back). This distance will determine the size of the cap (small, medium, or large). To place the cap, mark the 10% of the measured distance on the forehead and make sure that the mid-frontal electrode (FPz) is placed on this marked point.
- Attach the face electrodes to their respective cords on the cap
- Start filling the electrodes with gel, using the blunt needle tip to scrape the hair aside underneath the electrode, so the electrode is in direct contact with the scalp. Be mindful not to injure the skin.
- Lifting up the electrode a bit makes it easier to insert the gel. In most cases, there will be hair underneath the electrode. Moving it out of the way will allow for better impedance.
- Take the participant to the soundproof room and plug in the cap and individual electrodes.
- Check the impedance of the electrode-scalp connection to keep it under 10 KΩ. If the impedance is high make sure the electrode has conductive gel and is in touch with the scalp.
- Impedance is the tendency to impede the flow of an alternating current. High impedance may increase noise in the data, and should be minimized before the study begins.
- In most cases, the hair is in the way of the electrode. Moving it out of the way should get better impedance.
- Once the impedance is acceptable for all electrodes, and EEG traces are void of noise, data collection can begin.
Figure 1: Electrode placement. Placement of the face electrodes to detect EOG artifacts (left). Diagram of measurement from directly between the eyebrows to just under the bump in the back of the head. 10% of this measurement is measured above the mid-eye mark, and this is where the FPZ electrode of the cap is placed (right).
- EEG data collection
- Prepare the participant to do the task.
- Place the participant in a chair 75 cm from the 16-in. computer screen, in a sound and light-attenuated room (acoustically and electrically shielded).
- Tell the participant that he/she will be seeing colored circles appear on the screen. Every time a green circle is seen, the participant should press a button held in his/her right hand (Figure 2).
- Show each stimulus for 1000 ms, with a 1000 ms interstimulus interval between stimulus presentations.
- Show the 64 target stimuli, randomly interspersed among 96 presentations of the non-target red circles. Repeat this sequence twice, for a total of 128 target stimulus trials and 192 non-target control trials.
- Start the system, and have continuous recording of EEG throughout the presentation of the functional task.
- EEG is amplified by amplifiers with a gain of 1024 and a band-pass of 0.01-100 Hz.
- Trials contaminated by eye-blinks and artifact rejection (approximately 15% of trials) will be eliminated off-line.
- Prepare the participant to do the task.
Figure 2: Study design for the oddball task. The participant is presented with either a red circle or a green circle. Each stimulus appears for 1 s, followed by a 1-s blank screen. Each time the participant sees a green circle, he is instructed to press a button held in his right hand.
3. Data analysis
- Offline, reference data to averaged mastoids.
- Segment continuous EEG data into epochs, starting 200 ms before and ending 1000 ms after the onset of the stimulus.
- Epochs are baseline corrected using the epoch 200 ms before the onset of the stimulus.
- To correct for motion artifacts, epochs with a signal change exceeding 150 microvolt at any EEG electrode were not included in the average.
- The data are digitally filtered offline (bandpass 0.05-20 Hz).
- Use the ERP averages that are displayed from the Pz recording sites for target and control stimuli.
- The peak (amplitude and latency) of the parietal P300 is automatically obtained at electrode Pz.
- Statistical analysis
- Plot ERP averages from the parietal Pz electrodes.
- For peak amplitude and latencies, use F-tests for each latency range to determine whether there is a difference between target and control stimuli.
Given the overwhelming amount of sensory information in our environment, the brain must be able to prioritize the processing of certain stimuli, so that it spends less effort on what might not be currently important, and attend to what is.
Everyday, a person is exposed to multiple sights and sounds, like people typing in the office or visuals on a computer screen.
The attention someone pays to such stimuli depends, in part, on their goals at any given moment. For example, they may purposefully focus on their monitor to review a presentation. When this happens, the brain ignores frequent, unimportant items—like the typing of coworkers—and instead attends to the slides on the screen.
This is an example of a process called top-down attention, in which the brain filters out information not related to a goal.
In contrast, bottom-up attention deals with unique, unexpected stimuli, which have the ability to capture a person’s attention, even though they aren’t related to an objective.
These rare noises or sights are called oddball stimuli, and—due to their novelty—are prioritized by the brain for processing, as they may be important. A noise in the kitchen might mean that someone is hurt, or that there may be snacks.
In response to such important sensory events—the crash in the kitchen—multiple neurons in the same region of the brain can be activated, which promotes the propagation of an electrical signal.
This electrical response can be gauged at the scalp with electrodes through techniques of electroencephalography—abbreviated as EEG—and the resulting measure is called an event-related potential, or ERP.
In this video, we will investigate ERPs during an oddball paradigm, in which subjects are shown unique and common visual stimuli. We will demonstrate how to setup an EEG experiment, analyze ERP data, and explore how researchers are applying this technique to study other aspects of attention.
In this experiment, the brain activity of participants viewing two types of shape-based stimuli—baseline and oddball—is measured using EEG, in order to gain insight into how the brain identifies irrelevant from important sensory information.
To prepare for EEG, researchers position electrodes—already inserted into a cap—on participants’ scalps at specific anatomical locations, so that electrical activity in the brain can be recorded.
Additional electrodes are placed around the eyes to gauge muscle activity—which can produce motion artifacts in EEG data—and behind the ears at mastoid locations that serve as references where non-neural information is gathered.
Participants are then introduced to the two types of stimuli they’ll be seeing. Here, baseline visuals consist of a single, red circle, whereas an oddball image is composed of an individual green circle.
Participants are instructed to be on the lookout for green shapes, and directed to push a button whenever one is shown onscreen.
During the task, each circle appears for 1 s on a computer monitor. After the circle disappears, the screen remains blank for 1 s, and the next image is then presented.
The trick is that participants are shown green circles sporadically—and much less often—between several sequential images of red ones. Of the 160 stimuli, only 64 are green.
The idea is that these "out-of-place" target images will capture both bottom-up attention—as they are rare—and top-down attention—as the goal of the task is to indicate when these shapes appear.
As a result, the brain will respond to these goal-related, potentially important stimuli by producing robust electrical signals.
EEG data are continuously recorded over all 160 trials. Then, the sequence is repeated and a second set of 160 images is shown, which ensures that enough information is collected to distinguish true oddball-induced activity from noise.
Afterwards, EEG data are processed to generate ERP waveforms for each anatomical site over which an electrode is placed.
Based on previous research, the most critical data are expected near the Pz electrode, located in the center of the scalp towards the back of the head, above the junction of the parietal lobes.
Specifically, one component of these parietal ERPs, called P300—so named because it consists of a positive peak in the waveform that occurs approximately 300 ms after a sensory stimulus is presented—are predicted to be enhanced in response to green oddball circles.
To begin the experiment, greet the participant and ensure that they sign all appropriate consent forms. Also confirm that they have not used any hair products, such as mousse, which could interfere with EEG recordings.
Before proceeding, first stir the conductive electrode-gel to release any air bubbles. Then, use it to fill a 10 ml syringe, which will aid in applying this substance to electrode arrangements later in the protocol.
Once the syringe has been prepared, thoroughly brush the participant’s hair and scalp, and clean the top of their head with cotton gauze soaked in alcohol.
Afterwards, sterilize the skin behind each of the participant’s ears above and below the left eye and at the far horizontal positions of both eyes in a similar manner.
Next, place one face of a two-sided adhesive disk against an electrode. On the other side, apply gel onto the exposed electrode, and then affix it to the cleaned area above the left eye. Repeat this process at the remaining sterilized positions on the face.
To determine the size of the EEG cap to be used, measure the distance from the front of the participant’s head—directly between the eyebrows—to the inion projection of the skull, located below the bump at the back of the head. Above the mid-eye point, mark 10% of the measured distance on the forehead.
Using the eye-to-inion measurement, choose a cap that fits within standard circumference ranges, and place it so that the FPz electrode—the central, front-most one—is positioned over the mark on the forehead. Then, connect each of the face electrodes to its respective cord on the cap.
After retrieving the gel-filled syringe, inform the participant that you’ll be inserting the blunted tip into every electrode. Now, lift each one, and scrape the underlying hair aside, being careful not to injure the skin.
Then, proceed to insert the gel, and repeat this process for the remaining cap electrodes to ensure that the electrical signals collected at the scalp are properly conducted.
Afterwards, take the participant to a quiet room with acoustic and electrical shielding, and plug in the entire cap into the recording system.
Using the associated computer program, check the impedances of the electrode-scalp connections. If the impedance is above 10 KΩ for any electrode—which can result in noise in EEG traces—verify that it has conductive gel and that all of the underlying hair has been moved away.
Note that all impedance values should now be below 10 KΩ.
In preparation for the behavioral task, have the participant sit so that they are positioned approximately 75 cm from the monitor, and give them a response box to hold. Emphasize that they should only press the timing button when they observe a green circle onscreen.
After the participant understands the task, start the EEG system. Allow them to complete the 64 target stimulus trials interspersed with the 96 non-target control trials. Following the 160 trials, initiate the sequence to repeat.
Once all of the data have been collected, import the results into an analysis program to begin offline processing. First, isolate only the neural signals by referencing the information to averaged mastoid values.
Continue by dividing the continuous EEG recordings into epochs—sections beginning 200 ms before and ending 1000 ms after the onset of every stimulus, in this instance either green or red circles.
Proceed to baseline adjust these time frames using the portions that occur 200 ms prior to the stimulus onset.
Then, to correct for motion artifacts, eliminate periods during which a signal change exceeding ±150 µV was recorded at any of the electrodes—not just the one collecting data from an anatomical site of interest.
Afterwards, for each electrode, average the EEG data collected from all baseline image trials to produce an ERP waveform. Similarly, average the data for oddball trials.
To analyze the data, display the ERP averages from the Pz recording site for both the green target and red control shapes. For the parietal P300 components, assess the amplitude—the height of the component above the baseline value of 0 µV—and its latency—how long it appears in ms after the participant views the circle.
Then, for these peak amplitudes and latencies, use F-tests to determine whether there is a difference between baseline and oddball stimuli.
Notice that for green oddball shapes, the trace peaked at approximately 350 ms after the onset of the stimulus, whereas no P300 peak was observed when the participant viewed the control red target circles.
Collectively, these data suggest that activity in the parietal lobe increases when oddball stimuli are presented, reflecting the neural processes that identify task-relevant, salient stimuli.
Now that you know how researchers use the visual oddball paradigm in sensory information processing, let’s take a look at how scientists are applying this technique to analyze ERPs in other areas.
Although we’ve focused on ERPs produced by healthy brains, some researchers are using the oddball paradigm to understand how concussions—injuries resulting from trauma to the head—affect cognitive processes.
For example, evidence exists that students who suffered a concussion—and who exhibit symptoms such as dizziness or confusion—produce significantly lower P300 peaks when they view a rare oddball image, compared to uninjured control participants.
This suggests that concussions may negatively affect how the brain responds to, and processes, potentially important sensory information.
Other researchers have used a modification of the oddball paradigm—one involving unexpected sounds, rather than images—to better understand the differences between top-down and bottom-up attention.
By studying ERPs produced by oddball tones, scientists have determined that, not only is the P300 peak also enhanced by rare sounds, but that this component can consist of two subparts: an early portion called P3a, and a later P3b element.
Interestingly, these two peaks are observed in the ERPs of participants given the goal of identifying rare sound stimuli. However, only P3a occurs in the waveforms of participants told to just passively listen to sounds, and not given a goal to identify odd ones.
Thus, P3a is thought to deal with bottom-up attention—and how the brain responds to novel stimuli—whereas P3b likely reflects top down attention, and how the brain cognitively classifies targets.
You’ve just watched JoVE’s video on using the oddball paradigm to investigate the processing of sensory stimuli—particularly in the parietal lobe. By now, you should know how to design different stimuli, record EEGs, as well as generate and analyze ERPs. You should also have an understanding of how ERPs can provide insight into cognitive processes, and be used to better understand certain injuries.
Thanks for watching!
During the oddball task where participants were instructed to respond with a button press each time they saw a green circle, there was an increased parietal P300 compared to when the participant viewed the control red circle. This trace peaked approximately 350 ms following the onset of the stimulus, whereas there was no P300 peak for the control trace (Figure 3).
Figure 3: P300 parietal response to baseline and oddball images. Average ERP time trace of the parietal response to baseline images (red) and oddball images (green). The response is measured in microvolts over milliseconds.
These results show that activity in the parietal lobe increases when an oddball item is presented, reflecting the neural processes that identify task-relevant, salient stimuli. The brain increases its efficiency by identifying these items and focus resources on processing them. Stimuli which capture attention in this way are responded to more quickly, and also remembered better later.
Applications and Summary
The ERP approach, due its very high temporal resolution, allows discrimination between the electrical events that correspond to extremely fast psychological processes. The oddball task demonstrates this power, in revealing an electrical signature from the parietal lobe that discriminates between two similar stimuli less than half a second after their presentation. The task provides a window into the brain's process for identifying features in the environment that have current biological importance.3
The oddball paradigm combines aspects of both bottom-up and top-down attention. Bottom-up attention refers to the exogenous ability of a stimulus to capture our attention regardless of our own willful plans or goals. This comes into play in the oddball task in that the targets are rare and different from the other stimuli in the experiment, which makes them stand out. Top-down attention refers to our ability to filter incoming information based on our current task goals. The oddball task involves aspects of top-down attention because we are instructed to respond only to the target stimuli, therefore we are consciously trying to attend to them. Research has found that the P300 potential may have early and late subcomponents, the early subcomponent (called P3a) reflecting the bottom-up saliency that is driven by the novelty of the stimulus, and the later subcomponent (called P3b) that reflects the top-down cognitive classification of the stimulus as a target. The oddball task is therefore a robust and complex probe of attentional processes.
As a reliable marker of attentional processes in the brain, the P300 elicited by the oddball task can be a useful biomarker of attentional dysfunction. For example, children with ADHD show a smaller and later P300 potential,4 and these differences tend to decrease with effective drug therapy.5
- Squires, N.K., Squires, K.C. & Hillyard, S.A. Two varieties of long-latency positive waves evoked by unpredictable auditory stimuli in man. Electroencephalogr Clin Neurophysiol 38, 387-401 (1975).
- Habibi, A., Wirantana, V. & Starr, A. Cortical Activity during Perception of Musical Rhythm; Comparing Musicians and Non-musicians. Psychomusicology 24, 125-135 (2014).
- Halgren, E. & Marinkovic, K. Neurophysiological networks integrating human emotions. in The Cognitive Neurosciences (ed. Gazzaniga, M.S.) 1137-1151 (MIT Press, Cambridge, MA, 1995).
- Doyle, A.E., et al. Attention-deficit/hyperactivity disorder endophenotypes. Biol Psychiatry 57, 1324-1335 (2005).
- Winsberg, B.G., Javitt, D.C. & Silipo, G.S. Electrophysiological indices of information processing in methylphenidate responders. Biol Psychiatry 42, 434-445 (1997).