A Review of An Introduction to the Event-Related Potential Technique
Introduction
When a participant enters an EEG lab, they might have likely been recruited by a psychology department given that we are looking at the brain, which is the powerhouse behind our human behavior. However, the actual technology we are using, Electroencephalography (EEG), functions by laws of physics. While fMRI and PET technology is doing a great job at mapping out the structural and functional dimensions of the brain, it is important to understand the actual brain power electrical activity. We do this by measuring the voltage that is emitted from the cerebral cortex and picked up by the EEG cap electrodes.
Once the lab assistants have fit the EEG cap onto the participant, and have checked for good conduction, the participant is then briefed on what it is they will be doing. The tasks that the participant does always involves an event that stimulates an electrical response of the brain. This is known as an event-related potential (ERP) and it denotes the electrical potential that is related to specific events (Luck, 2014). The most amazing part of ERP research using EEG technology, is that it records the exact moment the current flows from it’s cortical Pyramidal cell. While EEG is very exact at recording when the EEG amplitude changes, there is still much work to be done to overcome the ambiguity of exactly where these changes are occuring within the cerebral cortex.
The Cerebral Cortex
The average surface area of the human cerebral cortex is about 2200 to 2400 cm x cm. It is the outermost layer of the brain that consists of gray matter and is divided into a left and a right hemisphere by the medial longitudinal fissure, that goes right down the center of the brain. At only a few millimeters thick, it connects with subcortical areas, which are the areas below its surface. The cerebral cortex has three different areas: sensory, motor, and association areas. Sensory areas work with processing sensory information by collaborating with the Primary somatosensory cortex, Primary visual cortex, Primary auditory cortex, Primary olfactory cortex, and the Primary gustatory cortex (Alem, 2011). Motor areas initiate us to move and to act and consist of the primary motor cortex, Premotor cortex, Broca’s area, and the frontal eye field (Alem, 2011). Association areas are spread throughout the cortex and work as a sort of middle man, integrating and interpreting sensory inputs from it’s primary area to it’s associated area (Alem, 2011). Due to the extensive folding of the cerebral cortex, the neurons that spark dipoles, which are an oriented flow of current, send these currents of electricity in many different directions, especially the neurons that are deeper within the valleys, or sulci of the cerebral cortex. Ribboning beneath thy gyri and sulci that give the cerebral cortex it’s topography, is a thin layer of grey matter. This grey matter ribbons beneath sulci that run deep even within fissures of the brain, these are also known as deep sulcus. In order for event-related potentials to be measurable, thousands of similarly oriented neurons need to be combined in order to create an electrical dipole.
During an EEG study, once the participant has begun the study, the EEG is continuously at work recording the brain’s constant activity. These are known as EEG oscillations, and they are the result of the constant variations in the pattern of postsynaptic potentials across the billions of neurons in the brain (Luck, 2014). The oscillatory nature of the brain’s activity is reflective of the feedback loops in the brain. While this unveils much of the live activity of the brain, it requires special mathematical and statistical techniques to measure the ongoing activity of the brain. Fourier Analysis sums together the different amplitudes, frequencies, and phases of the different waves that come from the different oscillatory patterns of the brain’s activity. Fourier Analysis is the mathematical translation of a signal in time and its frequency decomposition (Hoang, 2016). In the case of measuring a participant’s response to an event, or a stimulus, fourier analysis is used when averaging the event-related potential waveform. Unfortunately, outside noise can pollute the ERP waveform therefore making filtering an important part of interpreting data in an ERP study. The EEG voltage is usually under 100 microvolts, uV, so the signal from each electrode is usually amplified by a factor of 1,000-100,000 (Luck, 2014). Along with filtering, amplifying the EEG voltage to be readable is also a challenge when interpreting data from an EEG study.
Voltage at the Scalp
A more precise term for voltage, at least in regards to event-related potentials in EEG research, is electrical potential. This is because the potential for electrical current to flow from one place to another is time-locked to a stimulus. Voltage is defined as the pressure that pushes the electrical current through a conductor, and in this case, the conductors are the metal electrodes on the EEG cap. The electrodes are arranged according the the 10/20 system and have both metal and added saline solutions to lower impedance.
The 10/20 system names each electrode site according to its position of the general brain region, hemisphere, and distance from the brain’s midline. The first letter of the 10/20 system represents the brain region in which it is on top of such as, F for frontal, C for central, P for parietal, O for occipital, and T for temporal. There is also Fp, which represents the frontal pole. After the first letter in the 10/20 system, either an odd or even number indicates which hemisphere the electrode is in, as well as how far away it is from the midline. Odd numbers are for the left hemisphere and even numbers are for the right hemisphere. As the numbers increase, so does the distance from the midline. If the electrode is on the midline, a lowercase z takes place of the number to represent zero (Luck, 2014).
Electrical currents spread out across the conductor therefore, event-related potentials are picked up by electrodes all over the head. This requires interpolation based on what we know about physics and what we know, so far, about the physiology and electrical activity of the cerebral cortex. We know there is electrical activity occurring as thousands of polarity changes occur to create postsynaptic dendritic currents. We understand that voltage is a better measure to use as oppose to current, because it is easier to measure the magnitude of the neural response, in uV, at each electrode site, than it is to measure all of the volume of the current.
If current were water and a conductor was a water hose, than voltage would be the water pressure. This is why it is easier to measure voltage, more so than it is to measure current. Resistance is also important to consider when discussing the physics of event-related potentials in an EEG study. Resistance is what hinders voltage, so in the water-in-a-hose analogy, resistance would be perhaps that the water is frozen or evaporated (the composition of the substance). The water hose could be too long, too thin, or too wide (the length or diameter of the hose). In an EEG study, we override resistance, by increasing the voltage (Luck, 2014). We of course, can’t increase the brain power but we can amplify it as we are recording it and we can still accurately time-lock the activity to a stimulus as having either a positive (+) or a negative (-) potential. This essentially is the voltage we are measuring and due to the oscillatory nature of the brain’s activity, using current as a means of measurement would become quite perplexing. More importantly though, no water will flow out of the end of the hose unless there is sufficient water pressure. Also, just as water completely fills the entire conductor, so does electricity. This is essentially why there is so much guess work in measuring ERP components. When creating an interpolation temperature map of an area, meteorologists average the different temperatures of different regions of an area to present an approximation on the weather report. It is the collective whole of the electrodes all over the head that reflect the electrical responses of the participant’s brain in response to an event.
Event-Related Potentials
When recording the electrical potentials that are related to specific events, the either positive (+) or negative (-) potential is plotted on a Y-axis, while time (milliseconds) is plotted on the X-axis. This accounts for the oscillatory fashion of the electrical activity of the brain, and reflects the exactitude of recording how much electrical power the brain is giving off and at exactly what time this is occuring in response to a stimulus. To measure brain amplitude (or voltage), EEG is measuring the difference between two inputs, and all of the other commonalities are canceled out (Luck, 2014). Again, this is why it is easier to use voltage as a measurement as oppose to current, because it is easier to measure voltage difference between two points than it is to interfere with the actual current. Using current would be very noisy, and it would make isolating neural responses even more tricky than it already is.
We cancel noise through averaging positive (+) and negative (-) peaks to refine the data to focus on the signal of interest. While EEG technology is still technically in its infancy, especially with the even later developments that have improved our understanding of the spatial disposition of the brain, a major goal in cognitive neuroscience is to parallel the microvolts (uV) +/- peaks and valleys to a set of neural activity within the cerebral cortex. This first requires improving spatial resolution beyond our current improvements. Our current improvements include increasing the amount of electrodes we have a participant wear. The more electrodes a participants has on, the better an estimation we can make of exactly how much brain activity is actually being produced. Also using fMRI for it’s spatial resolution can improve your approximation of which cognitive processes are associated with the peaks (+) and valleys (-) that are timed against an event.
Going back to the topography of the cerebral cortex, two of the main constrictions on spatial resolution in an EEG study is the topography of the cerebral cortex, or the inverse problem, and the ambiguity of the underlying subcortical components that is hushed by the layers of white and gray matter, skull, scalp, and hair that lie above it, which is also known as the superposition problem. The inverse problem states that there are multiple different sets of dipoles that can perfectly explain a given voltage distribution, and there is no way to tell which is the correct one (Luck, 2014). Due to the inverse problem, action potentials are not able to be detected at the scalp. The voltage from postsynaptic potentials is recorded at the scalp, however. Event-related potentials originate as postsynaptic potentials (PSPs), which occur when neurotransmitters bind to receptors, which then changes the flow of ions across the cell membrane. Scalp ERPs arise from pyramidal cells, which are the messenger cells of the cerebral cortex, communicating between the cerebral cortex and it’s subcortical components that lie beneath its surface. Once these postsynaptic potentials generate an electrical dipole, a magnetic field can be detected in a similar way as voltage. If we pay closer attention to these magnetic fields that engulf the event-related potential dipole, we would have improved spatial resolution than is possible with electrical potentials due to the transparency that the skull has to magnetism (Luck, 2014). The superposition problem is one that is unsuspecting because on the flip side, it reflects a major strength of EEG, which is it’s temporal resolution of 1 millisecond or better. This is significant, especially in comparison to fMRI or PET technology which takes several of hundreds of millisecond due to it needing to wait for the blood to travel through the body. It’s fast, but it’s so fast that it’s difficult to decompose the mixture into the individual underlying components (Luck, 2014).
Cognitive Processes and Their Components
P3
Another name for the P3 component is the P300 component, which is P for a positive uV, and 300 for 300 ms from onset of stimulus. Unexpected unusual, or surprising task-irrelevant stimulus within an attended stimulus train will elicit a frontal P3-like response (Luck, 2014). However, when a subject is uncertain of what is a target, the P3 amplitude would be minimized. The P3 component ultimately helps you prepare for the future, but occurs too late to have an impact on the behavioral response.
N170
A negative component that peaks around 170 ms after stimulus onset is notable because it generates a larger stimulus response to when the subject see’s a face compared to when they see a non-face object.