This post is the second in a series that aims to educate readers about the tools that are used in neuroscience research. Last month we discussed Calcium Imaging.
Electroencephalography (EEG) is a powerful technique for assessing brain activity. It has been around much longer than newfangled methods like Positron Emission Tomography (PET) and functional Magnetic Resonance Imaging (fMRI), but remains relevant today for purposes as diverse as epilepsy treatment, prosthetics, and biofeedback.
Many people have heard of EEG in the context of “brain waves” and the different frequencies of brain activity associated with sleep and wakefulness. These different frequencies can be seen at right and the traces offer a nice glimpse of what raw EEG data looks like. Frequency analyses remain an important part of EEG research, but the technique is also capable of answering more complex questions about the brain’s thoughts and perceptions.
In 1924, the German doctor Hans Berger coined the term electroencephalography after making the first voltage recordings across the human scalp. Berger discovered an 8-12 Hz rhythm that was present when subjects relaxed but disappeared when they opened their eyes, the alpha wave. As years passed and research progressed, the electrophysiological correlates of other cognitive states were also elucidated. Before discussing these advancements, let us examine how the electrical signals are generated in the first place.
When neurotransmitters are released at a synapse, they cause ion channels to open on the postsynaptic terminal of the next neuron. This results in an influx of positive ions that depolarizes the neuron, also known as an excitatory post-synaptic potential (EPSP). The local extracellular environment, depleted of positive ions, takes on a negative charge. As this current propogates down the conductive dendrite of the neuron, the size of the EPSP decreases. This means that the net depletion of positive ions is greatest outside the synapse and smallest outside the cell body, setting up an extracellular voltage difference along the axis of the neuron. This extracellular voltage difference represents the sum of the neuron’s inputs instead of its output (e.g. action potentials). Every neuron receiving synaptic inputs can therefore be thought of as a dipole with a specific orientation and polarity. A dipole corresponding to a single neuron is not detectable with EEG, but when thousands of neurons with similar orientation receive similar synaptic inputs, the dipoles sum together to yield strong voltage signals at the scalp.
Scalp voltages can be measured using a cap studded with electrodes. Sometimes a conductive gel is applied between the scalp and the electrode to improve the signal. The voltage at each electrode is compared to a reference electrode and amplified to increase the signal to noise ratio. Then the signal is sent to a computer where it can be filtered and amplified.
Unfortunately, a given distribution of scalp voltages can be produced by an infinite number of unique dipole arrangements. For instance, two nearby dipoles of opposite orientation might cancel out (this phenomenon is actually quite common because the cortex contains many folds). This makes it very difficult to draw conclusions about the neuroanatomical regions that are involved in a given EEG signal.
Although EEG has much poorer spatial resolution than newer techniques like fMRI, its real-time access to electrical activity provides vastly superior temporal resolution. This property makes it ideal for investigating stereotyped neural responses to stimuli. The corresponding EEG recording is called an Event-Related Potential (ERP). Often recordings from a single stimulus presentation will bear little resemblence to the expected waveform. This is because the responses to many stimulus presentations must be averaged together to yield clean data.
Here is the average ERP waveform in response to sequence of auditory stimuli. The majority of stimuli are identical, while occasionally the subject is presented with a deviant stimulus of a different tone. The gray trace corresponds to the average response to standard stimuli whereas the black trace corresponds to the average response to novel stimuli. It is marked by an increased positive potential at around 300 milliseconds, termed the P300. Because EEG signals are sensitive to factors like uncertainty and probability, they are thought to contain information about higher order cognition.
ERPs are also useful in a clinical setting. Visual evoked potentials can be collected in response to a flashing checkerboard to examine whether a patient’s visual cortex is functioning properly. EEG recordings are also useful for locating epileptic seizures in the brain. Sometimes surgeons will even implant electrodes beneath the scalp or inside the brain to get a more accurate picture of local brain activity. Incidentally, this was how Halle Berry neurons were first discovered.
Interested in learning more about EEG?
- Open EEG is an online community for electronic hobbyists interested in building a homemade setup. They provide the plans and advertise that a complete system can be built for around $250 dollars.
- Because EEG has high temporal resolution and fMRI has high spatial resolution, scientists are increasingly interested in combining the techniques. You can find a summary of this approach here.
- Recently the theory of mirror neuron dysfunction in autism has received a lot of hype. EEG recordings provided the first evidence for this theory.
- Researchers are also hoping to control prosthetic limbs with EEG.
- Lastly, EEG is increasing in popularity as a tool for biofeedback. Here is a news clip about the fad.