Neurophysiology Background: What is ERP?
Background to the ERP
Buried in the EEG recorded specifically during the brain's response to any internal or external events, is a signal that is much more revealing about human information processing. This signal is called the Event-Related Potential (ERP). In the 1950s, two decades after the discovery of EEG, a technique was developed in order to isolate this ERP signal from the EEG. This was one of the most significant advances in cognitive neuroscience. The ERP is the portion of the ongoing EEG pattern of the brain that is accurately time locked to a particular occurrence. It is so called because it is the change in 'potential' or voltage that is 'related' to a specific 'event'. This 'event' may be a sensory stimulus, a cognitive event, or the execution of a motor response.
The synchronous nature of the ERP is its fundamental strength and represents its major advantage over the traditional EEG measure. EEG activity reflects a wide range of ongoing neural activities related to the myriad self-regulating systems, sensory functions and cognitive operations. This intermixing of signals makes it difficult to separate out one factor from another. In contrast, because the ERP is time-locked to the onset of a specific internal or external event, the actual relationship between the neuroelectrical response and the event of interest is revealed. This relationship can be resolved down to the milliseconds time scale. Thus, ERPs can be regarded as 'manifestations' of stages of information processing in the brain. They reflect the changes in ongoing brain activity over time, which occur in preparation for, or in response to, discrete sensory, cognitive or motor events. Such changes are observed as differences in the voltage of the wave at various points in its time course, or as differences in the latency of particular peaks within the wave. The ERP, with its high temporal resolution, is now one of the most widely used tools for assessing various aspects of information processing, particularly in clinical research.
Generation of the ERP
ERP is generally believed to reflect the summed post-synaptic dendritic potentials of sizable populations of synchronously activated neurones, aligned in a parallel orientation in cortical or subcortical regions. However, the eventual ERP recorded at the scalp is by no means always an accurate and reliable reflection of all the discrete neural events related to the stimulus and is in fact a composite of various factors, only some of which relate directly or indirectly to the variables being manipulated in the experiment. ERP amplitudes tend to be low, ranging from less than a microvolt to several microvolts, compared to tens of microvolts for EEG, millivolts for EMG, and often close to a volt for EKG.
Recording of the ERP
Electrode Placement
Initially, in the 1930s, ERPs were recorded by applying electrode plates directly on the scalp. However, one problem with this recording technique was that the very minute ERP signals were often distorted or obscured by electrode movement artefacts. This was overcome with the introduction of a floater type electrolyte that required an electrolyte paste to be applied between the electrode and skin. The electrolyte allowed the small ERP currents to be more easily transferred to the electrodes. All the electrodes in use today are based upon this electrode type. Typical ERP studies place electrodes over bilateral frontal, temporal, central, parietal and occipital areas of the brain. This placement provides information relating to left versus right hemisphere responses to the evoking stimuli, as well as information within each hemisphere concerning the functioning of different brain areas. Varieties of strategies have been used to select electrode placements sites. Approximately half of the ERP studies in current literature use the 10-20 System designed for use with adults and children in which the location of the electrode is specified in terms of its proximity to particular brain regions and its hemispheric location (Jasper, 1958). The use of a standard system enables between-laboratory and between-experiment comparisons to be made. In recent years, an increased interest in the precise functional and anatomical generators of activity in the brain has necessitated the sampling of electrical fields at a higher spatial frequency. Accordingly, the 10-20 system has been enhanced by the use of non-standard locations and a higher density of electrodes. The Geodesic Sensor Net and the Electrocap are two such electrode systems.
Reference and Eyeblink Electrodes
The ERP activity is usually recorded using a 'common reference' recording procedure, which involves connecting each of an array of recording electrodes to single reference, comprising one or a pair of electrodes linked together. The reference site is chosen because it is uninfluenced by the electrical activity of experimental interest. The recordings are then based on the difference in voltage between each 'exploring' electrode and the same common 'reference' electrodes. Additional electrodes are usually placed in fixed positions at supraorbital (above each eye), suborbital (below each eye) and canthal (to the side of each eye) positions to enable the detection of artefacts due to horizontal and vertical eye movements and eye-blinks.
Extraction of the ERP
Sampling
The ERP recording of each experimental trial comprises a number of samples taken at fixed intervals throughout the sampling period. Sampling is necessary because the brainwaves are analogue (continuous) signals and ERP analysis can only be performed on digital (sampled) signals. The sampling period and sampling frequency are usually based on previous studies investigating similar phenomenon, or on studies of similar populations.
Averaging
One fundamental difficulty in extracting the ERP is that the small ERP (~5-10 _V) 'signal', reflecting the relevant neural event, is obscured by the larger EEG (~50 _V) 'noise', generated by the myriad ongoing neural processes. In the 1940s, Dawson devised a technique to improve this low 'signal to noise' ratio, using a capacitance based computer analogue that repetitively summed up and averaged the elicited ERP signals over a number of repeated trials. These averaged ERPs included the information common to all the ERPs collected during the recording session that had a constant temporal relationship to the event. That is, the averaged ERPs reflected the relevant, repetitive time locked neural activity, while the non-repetitive signals reflecting random activity failed to contribute systematically to these specific portions of the ERP average. All the modern ERP systems are extensions of this original averaging design.
Filtering
Another technique to remove artefactual electrical activity from sources other than the brain and to improve the signal to noise ratio, is to filter the ERP in the frequency bands of interest. Usually the amplifiers that record the ERP include filter settings that eliminate any activity above and below selected frequencies. This allows the attenuation of very high-frequency electrical activity, such as the activity that is attributable to muscle rather than brain, or the activity at the electrical mains (50-60 Hz). The ERP signal is filtered further with analogue or digital filters after it has been averaged. This helps to increase the signal to noise ratio, since it attenuates the EEG frequencies that are either higher or lower than those contained in the ERP signal of interest. However, this is possible only if the frequency of the signal and noise are sufficiently different.
Removing Artefacts
Two major sources of artefact, eye movements and blinks, cannot be filtered out since these movements occur at the same frequencies as key features of the ERP signal. These movements produce fluctuating electrical fields that are consequently picked up by scalp electrodes, and contaminate the recordings of relevant brain activity. A common procedure to deal with these artefacts is to use the eye movement recording electrodes (described previously) to estimate and remove the contribution of these movements to the ERP signal. Correcting the ERP signal in this manner allows all recorded ERP data from other electrodes to be retained even if eye artefacts are present.
Definition of ERP Components
In any ERP research, after recording and extracting the ERP, the focus of interest turns to particular features or components of the resulting waveform. However, the human ERP reflects a complex and interactive process between an individual and the environment. Consequently, the question of what defines an ERP component has caused much controversy. Several parameters define and distinguish between ERP components, and the interpretation of the ERP data as well as the experimental or analytical methods in ERP research depend, in part, upon the particular parameter that is chosen. In general, three defining characteristics for ERP components have been proposed.
Waveform Based Definition
Here, the defining characteristic is the ordinal or temporal relationship between the components (peaks and troughs) in the waveform. The peaks can be named according to their polarity and sequence, which can be in numerical or alphabetical order. The 'fast' components are a series of small positive waves recorded in the first 10 ms, and numbered sequentially using Roman numerals. The 'middle' components occur between 10 and 50 ms, and are identified by their polarity and by a combined numerical and alphabetical sequence. The 'slow' components occur between 50 and 500 ms, and are identified as P1, N1, P2 and N2. However, such a sequential nomenclature is often the cause of some confusion and complexity. The peaks can also be named according to their polarity and latency to peak, in milliseconds from stimulus onset. Thus, in this nomenclature the fast wave 'V' of the auditory response would be called 'P6' and the slow component 'N1' would be identified as 'N90'. This nomenclature is more simple and accurate than the sequential nomenclature but its very precision can sometimes cause problems, since the latency of a component may vary with the intensity or rise time of the stimulus, or with the time required for perceptual processing. For instance, wave 'V' varies in latency from 5.5 to 8.5 ms as intensity decreases, without changing its identity, and the late positive component, which is related to perceptual decisions, may vary in latency from 300 to 900 ms depending on decision difficulty. In most ERP literature, the 'slow' components of the signal are identified using the 'sequential' nomenclature, to emphasize the function-related sequence, rather than accuracy-based latency of the component.
Source Based Definition
According to this view, the defining characteristic is the particular anatomical source (brain structure or system) of the particular component. Thus, the key parameter is the location of recording electrodes. With this defining criterion, it is quite possible that components with different polarity and latency recorded from different regions reflect the same underlying component. For instance, part of the primary cortical somatosensory response is generated in the posterior wall of the Rolandic fissure, creating a dipole that is parallel to the scalp surface. The same ERP component can therefore be recorded from the anterior scalp regions as a 'P20' wave and more posterior regions as a 'N20' wave. Conversely, it is quite possible that components with the same polarity and latency, recorded from different scalp areas, reflect separate and distinct underlying components. For instance, in response to visual stimuli, the 'N150' recorded in from the vertex and the 'N150' recorded from the occipital regions are generated by different underlying systems. Given the difficulty in precisely isolating the source of the signal in any given electrode, this defining characteristic is perhaps the most controversial.
Process Based Definition
According to this view, the defining characteristic is the particular functional source (sensory, perceptual or cognitive processing) associated with that particular component. Thus, components are defined in terms of the cognitive function performed by the brain structures or systems whose activity is being recorded at the surface and this function is specified by the nature of independent variables whose manipulation affects the component (such as stimulus features or task demands) as well as by the relationship observed between the component and other measures (such as response time and accuracy). Process based labels are assigned to components with posited functional roles, such as "mismatch negativity", "contingent negative variation", or "readiness potential". The defining characteristic is therefore the antecedent conditions and experimental manipulations associated with the component. This is arguably the most useful means of distinguishing ERP components at present. This classification usually depends upon the relationship of the component to external stimuli. According to this stimuli-based criterion, ERPs can be classified as evoked or emitted or as exogenous and endogenous.
Evoked and Emitted ERPs
'Evoked potentials' are the ERPs that follow a physical stimulus whereas the 'emitted potentials' occur in the absence of any evoking stimulus. Evoked potentials will be discussed in more detail later in this section. Emitted potentials may be associated with some psychological process such as the recognition that a stimulus has been omitted from a regular train, or with some preparation for an upcoming perceptual or motor act. The evoked potentials can be further divided into 'transient', 'sustained' and 'steady state' responses. Transient potentials are elicited by a stimulus change whereas sustained potentials occur during the continuation of a stimulus. Steady state potentials are evoked by stimuli of sufficiently high repetition rate (stimuli presented in rapid succession) such that there is an overlapping of transient responses to form a continuous waveform with constant amplitude and phase relationship to the repeating stimulus. This wave might be buried in the ongoing EEG activity, but can be detected by Fourier analysis, since the frequency of the wave corresponds to the stimulation rate. The emitted potentials can be further classified into 'preparatory' and 'integrative' potentials. The preparatory potentials are those related to specific motor activity such as the 'readiness potential' that occurs before self-paced motor actions. The integrative potentials are related to complex human perceptual activity, and include waveforms such as the contingent negative variation.
Exogenous and Endogenous ERP Components
ERP components can be classified as exogenous or endogenous. 'Exogenous' ERPs are determined by the physical characteristics of a stimulus whereas 'endogenous' ERPs are determined by the psychological or cognitive significance of the stimulus, that is, the psychological or cognitive demands of the situation. Evoked potentials can have both exogenous and endogenous components. Emitted potentials, being related to psychological processes rather than physical stimuli comprise, by definition, have endogenous components only.
The Combined Approach to Component Definition
Although it is easy to describe the physiological and psychological approaches to component definition as though they are mutually exclusive, it is important to note that both approaches play an important role in the definition of an ERP component. An important distinction must always be made between the 'observational' terminology, which refers to the observed waveform features in a given ERP trace and the 'theoretical' terminology, which designates the ERP components according to the functional processes that they may represent. In what is now considered a classic and conventional approach to component definition, the combined approach defines a component by a combination of its polarity, its characteristic sequence or latency, its distribution across the scalp, and its sensitivity to antecedent conditions and experimental manipulations. Here, polarity and distribution are 'observed terms', implying a consistency in anatomical and physiological source, while latency/sequence and sensitivity are 'theoretical terms', implying a consistency in psychological function.
Compendium of Event-Related Potential Components
Following the convention in most ERP literature, the main ERP components of the evoked potential will be described as per the 'process-based' parameter. However, this distinction is too broad and simplified to provide an accurate representation. Early sensory, exogenous components can be altered by cognitive manipulations (such as top-down effects of attention), while later endogenous, cognitive components can be altered by the features of eliciting stimuli, (such as stimulus modality). Therefore, it is more reasonable to describe the components within a 'spectrum' from exogenous to endogenous and that is coextensive with time. In this manner, the components occurring within the first 100 ms after stimulus onset are mostly exogenous, while those occurring after 100 ms are more endogenous. The components are presented here in increasing order of their latencies and, as such, in increasing sensitivity to cognitive factors.
Exogenous ERP Components
The deflections occurring in the first 100 ms of the ERP trace are related to the transmission of sensory information from the peripheral sensory system to the cortex and/or the arrival of the sensory information in the primary cortex. Indeed, all parts of a sensory system, from the receptor cell to the association areas of the cerebral cortex, can contribute to the sensory ERPs. In terms of the stages of information processing, these components could be thought to represent the early descriptive processes of perception. These are called the 'exogenous' components of the ERP trace. These sensory ERPs, being related only to physical stimuli and not psychological processes are, by definition, evoked potentials. As mentioned previously, their essential characteristics, such as amplitude, latency and distribution, appear to depend on the physical properties of the sensory stimulus, such as its modality and intensity, and are unrelated to any 'cognitive manipulations'.
Auditory Exogenous Components
In the auditory modality, the earliest components have extremely short latencies and are likely to represent the transmission of sensory information in the peripheral pathways. The amplitude and latency of these components depends upon the intensity and rate of presentation of the auditory stimulus. The 'fast' auditory components (I-V), recorded in the first 10 ms after an abrupt auditory stimulus (such as a click), represent the activation of the cochlear nerve and brainstem nuclei. The more complex 'middle' components, recorded between 10 to 50 ms, reflect auditory activity in the thalamus and cortex. Finally, the 'slow' components of the auditory response, recorded after the first 50 ms, reflect the activation of various areas of the primary and association cortex. These include the P1 component.
Visual Exogenous Components
In the visual modality, the earliest components have longer latencies than those in the auditory modality, possibly because the neurones in the main subcortical relay nucleus (the LGN) are configured in such a manner as to create 'closed' fields and consequently their activity is not recordable at the scalp. The earliest visual components, recorded from around 25 ms after an abrupt visual stimulus (such as a flash), represent activation in the occipital cortex. The amplitude and latency of subsequent components in the 50-150 ms range, depends upon the pattern, contrast and intensity of the eliciting stimulus. A distinct 'slow' deflection in the 80-120 range usually represents the P1 component.
Mesogenous ERP Components
Following sensory activation across all modalities, a large complex of deflections is recorded from the vertex region with latencies ranging from 100 to 250 ms, which are often referred to as the 'vertex potential' components. Their morphologies are similar across sensory modalities, suggesting they reflect comparable underlying processes in each sensory system, or derive from generators activated by all modalities. Thus, they might reflect the convergence of sensory information from different modalities onto areas of association cortex, particularly in the frontal lobe. It is believed several underlying sources contribute to the deflections in this latency range. However, it is still not fully known if there are distinct components that are related to stimulus features and others to cognitive factors. If there are such overlapping exogenous and endogenous components, they might reflect a general comparative process whereby incoming 'exogenous' information is related to relevant 'endogenous' memories before its interpretation. In terms of the stages of information processing, these components could be thought to represent the various integrative and inferential processes of perception and early attention. These are sometimes referred to as 'mesogenous' components, since they lie at the interface of purely exogenous and endogenous components (Picton, 1980). The three main mesogenous components are the N1, P2 and N2. These are all differentially affected by experimental manipulations and have different scalp topographies, which suggests that they are functionally independent and have different intra-cranial generators.
The N1 Component
The N1 is a negative deflection, recorded mainly over the frontocentral and temporal areas in response to auditory stimuli and the occipito-temporal areas in response to visual stimuli, with a latency that ranges from 50 to 200 ms after stimulus onset depending upon stimulus modality.
The MMN Component
In 1978, Näätänen first described the Mismatch Negativity (MMN) component of the ERP trace (Näätänen et. al., 1978). Since then, its application in the research of auditory processing disorders has rapidly escalated to make it one of the most intensively studied ERP components. In adults, the MMN is a negative potential over the frontocentral scalp and a positive potential beneath the mastoids, with a latency that varies from 50 to 250 ms after stimulus onset (Näätänen et. al., 1978).
The MMN is usually elicited using a 'passive oddball' paradigm. The 'oddball' paradigm involves the presentation of infrequent 'deviant' stimuli randomly interspersed among frequent, 'standard' stimuli. The deviant can differ from the standard in terms of a physical feature, be the absence of a stimulus among a train of regularly spaced stimuli, or be a change in the inter-stimulus interval among a train of regularly spaced stimuli. The term 'passive' refers to the fact that the subject can ignore the stimuli and is not required to consciously attend to or respond to them. In the auditory modality, the MMN is not only elicited by changes in physical features of a stimulus, such as its frequency, intensity, or duration, but change in stimulus patterns, changes in speech stimuli or changes in spatial location of a sounds. The MMN is only elicited by the deviant, and not the standard, stimulus. Thus, it is often displayed as the difference wave calculated by subtracting the standard from the deviant wave. The negative deflection from 100-300 ms then reflects the MMN component. The MMN occurs when there is a change in a repetitive sequence of stimuli, and thus appears to be a sensitive index of the perceptual detection of 'stimulus change'. It is considered to index the functioning of short duration sensory memory comparison process. That is to say, the MMN reflects the detection of mismatch between a stored sensory memory representation of the permanent and temporal features of the preceding standard, and those of the incoming deviant.
Two aspects of MMN are extremely relevant, and have made this the most widely tested mesogenous component of ERP, with many clinical and research applications. Firstly, the MMN has been obtained in response to changes in a number of physical features of the auditory stimulus including frequency, intensity and duration, as well as spatial and phonemic changes. Moreover, it is extremely sensitive to very fine stimulus differences, occurring even when these differences approach the psychophysical thresholds for discrimination. For instance, it has been found to occur when the difference between the standard and deviant stimuli is as small as 8 Hz or 5 dB. Stimuli that are more deviant generally elicit a more robust MMN. Thus, MMN provides a neuronal representation of the discrimination of numerous stimulus attributes. Secondly, the MMN is elicited passively, so it does not require attention or a behavioural response in order to occur. Indeed, it has been obtained during sleep in infants and adults, and during sleep and anaesthesia in animal models. Hence, its clear advantage in comparison with behavioural methods is that it can be recorded without the subject's active involvement in the task and therefore, it provides a direct means for measuring auditory discrimination accuracy without task related factors such as attention or response strategies. Overall, the MMN provides an accurate and objective neuronal representation of automatic, preattentive, stimulus discrimination and it can therefore be applied to the assessment of auditory processing disorders. Various neuropsychiatric and neurodevelopmental disorders are accompanied by abnormal MMN to deviant stimuli. For example, the temporal component of the MMN is attenuated, and its latency is greater, in autistic children with temporal lobe impairment and the frontal component of the MMN is attenuated in subjects with ADHD, dyslexia and schizophrenia.
The P2 Component
The P2 is a positive potential over the frontocentral regions of the scalp, with a latency that ranges from 150 to 300 ms after stimulus onset, depending upon stimulus modality and stimulus parameters.
Endogenous ERP Components
After about 200 ms, there are a large number of components related to the context, rather than the features, of the stimulus (Picton 1980). In terms of the stages of information processing, these components of the ERP trace could be thought to represent high-level cognitive processes, requiring conscious attention. These are called the 'endogenous' components of the ERP trace. As mentioned previously, their essential characteristics, such as amplitude, latency and distribution, appear to depend on the subject's interactions with the stimulus, such as attention and task relevance, and are unrelated to the physical attributes of the eliciting stimulus. The two main endogenous components of interest in this thesis are the N2 and P3 components.
The N2 Component
The N2 is a negative potential over the frontocentral regions of the scalp, with a latency that ranges from 150 to 400 ms, depending upon stimulus modality and scalp location.
The P3 Component
In 1965, Sutton and his colleagues first reported the P3 component in the visual and auditory modalities. It is now the most prominent and extensively studied 'endogenous' component in the prototypical ERP trace, partly due to its large size and partly due to the ease with which it can be elicited. This is an 'emitted' endogenous potential: it indexes purely cognitive processes and follows a point in time when task-relevant information becomes available to the subject, regardless of whether that point is related to the presence of a stimulus. The P3 is a positive potential with maximal amplitudes over the midline centroparietal regions of the scalp and with a latency that varies from 250 to 600 ms, depending on the stimulus modality and task parameters. The P3 component is relatively unaffected by stimulus characteristics. Thus, it can be elicited using stimuli from any sensory modality. In descending order of clinical use, the modalities are auditory, visual, somatosensory, olfactory, and taste stimulation. However, the manipulation of physical aspects of the stimulus can sometimes affect the P3. The latency and scalp distribution of the P3 wave differs with each sensory modality, although the amplitude of the P3 wave response is similar across modalities. For example, the P3 elicited by visual stimuli has a marginally posterior distribution and a slightly longer latency than auditory stimuli. This suggests that some sources generating the P3 may differ, depending on the modality of the eliciting stimulus.
The stimulus presentation protocol most commonly used to obtain the P3 component is the 'active oddball' paradigm. As described earlier, the oddball paradigm involves the presentation of unexpected or infrequent 'target' stimuli, randomly interspersed among frequent 'standard' stimuli. During the standard presentation stream, the target can represent a change in its physical features, the absence of its presentation, or a change in its ISI. The term 'active' refers to the fact that the subject is required to consciously attend to and respond to the stimulus. The most commonly used response task, to augment the P3, is the 'selective choice reaction' task. This is generally used in conjunction with the oddball paradigm, and it requires selective attention and response to the task-relevant 'target' stimuli, while ignoring the irrelevant 'standard' stimuli. The response can be a motor response, such as pressing a button, or a mental response, involving counting to target stimuli. Like the MMN component, the P3 is only observed following the target, but not the standard, stimulus. However, unlike the MMN component, the P3 does not occur in the absence of attention, so when subjects are asked to ignore the stimuli no P3 is elicited. It is important to note that two factors, stimulus infrequency/unexpectedness and task relevance, operate independently to modulate the P3 component. In fact, there is evidence that these factors produce different P3 components. The 'P3a' component that was first reported by Squires and his colleagues is best elicited by an unexpected or novel stimulus (for example, if a third novel event is introduced into the oddball task), is independent of task relevance, has an earlier latency, and has a more frontal scalp distribution. In comparison, the classic P3 (which is often referred to as the 'P3b' component to distinguish it from the P3a), is best elicited by attending to task relevant stimulus, has a later latency and has a more parietal distribution.
The most established view on the significance of the P3 component is that the P3 reflects the "updating" of working memory when stimulus events require the subject's model of the environment or context must be revised. The extent to which the updating process is activated and, hence, the amplitude of the P3 component depends upon the significance or relevance of the eliciting stimulus. The working memory for the standard stimulus is well formed, because of its frequent presentation. Conversely, the memory for the infrequent stimulus is less well formed. Upon detection of the rare stimulus, the memory representation needs to be revised or updated. It is crucial to point out that the P3 does not reflect the actual processes involved in the detection of the rare stimulus, rather it is elicited as a consequence of this detection. Thus, infrequent 'target' stimuli elicit large P3 components, since the immediate memory for the preceding target stimulus has decayed and is refreshed following the presentation of a new target stimulus. Conversely, frequent 'standard' stimuli yield smaller P3 components because they maintain stronger representations in the working memory and therefore do not require as much updating. Although the context-updating hypothesis has been an influential view on the significance of the P3 component, it has been recently criticized, and a second, much broader alternative has recently been proposed, which suggests that the P3 component reflects the allocation of resources and the amount of processing involved in the evaluation and categorization of the stimulus. Evidence to support this theory comes from reports that the amplitude of the P3 is inversely related to task difficulty and processing demands and that the latency of the P3 is inversely related to the ease of categorization. For this reason, the P3 latency is considered a measure of 'stimulus evaluation time' independent of any response selection processes.
Clinical Significance of Evoked Potentials
Evoked potentials are the event-related potentials that follow a physical stimulus. However the term is usually reserved for those stimulus-elicited ERPs that involve either stimulation of, or recording from the cerebral cortex, brainstem, spinal cord and peripheral nerves. Thus SNAP (sensory nerve action potentials) or CMAP (compound motor action potentials) as used in nerve conduction studies are generally not thought of as evoked potentials, though they do meet the above definition. Evoked potentials are usually induced from one sensory mode at a time. Visual evoked potentials may be induced by a flashing light, auditory evoked potentials by the onset of a sound, and somatosensory evoked potentials by a solenoid pulsing against a person's skin. A visual evoked potential (VEP) is the potential caused by the sensory stimulation of a subject's visual field. Although it usually refers to responses recorded from the occipital cortex, sometimes the term 'visual evoked cortical potential' (VECP) is used, to distinguish it from retinal or subcortical potentials. The multifocal VEP is used to record separate responses for visual field locations. Commonly used visual stimuli are flashing lights or checkerboards on a video screen that flicker between black on white to white on black (counter-phase flicker). Visual evoked potentials are very useful in detecting blindness in patients that cannot communicate, such as babies: if repeated stimulation of the visual field causes no changes in EEG potentials, this is probably because the subject's brain is not receiving any signals from his/her eyes. VEPs are also useful in the diagnosis of optic neuritis, which causes the signal to be delayed. VEPs are also used in the investigation of basic functions of visual perception.
Sensory evoked potentials have been widely used in clinical medicine since the 1970s, including SSEP (somatosensory), VEP (visual) and BSEP (brainstem auditory) evoked potentials. SSEP are elicited by an electrical shock to a peripheral nerve as with NCV; VEP by an alternating checkerboard stimulus (PSVEP, pattern shift visual evoked potential) or a flash (strobe light or LED), and auditory evoked potentials by a click or tone, usually presented through earphones. Gustatory, olfactory, and even nociceptive (pain-evoked) potentials also exist, though none of these have found widespread clinical application. In addition, transcranial motor evoked potentials, recorded either epidurally or from muscles following electrical or magnetic transcranial stimulation of motor cortex, have become increasingly important in recent years. BSEP originally were very useful for detection of brainstem tumors and acoustic neuromas; when clinical averaging systems were first developed, the BSEP was more sensitive than early CT scanners. After several generations of progress in clinical imaging, however, this application is of marginal usefulness. BSEP is still used as a method of screening hearing in neonates, who can not be effectively tested by behavioral audiometry (though evoked oto-acoustic emissions recordings may eventually supplant BSEP in this application). BSEP is also used for intraoperative neurophysiologic monitoring in skull base surgery such as acoustic neuroma resection, where the auditory nerve is at risk. All three of the commonly used sensory modalities are useful in clinical diagnosis and monitoring of MS (multiple sclerosis); this is the main application for PSVEP. SSEP is also useful in coma. SSEP and MEP (motor evoked potentials), along with EEG and EMG, are extremely useful in intraoperative neurophysiologic monitoring for a wide variety of neurosurgical, otologic, spinal and other surgical procedures.
Effects of EEG States on ERP Components
It is worth noting the role of the brain states (EEG frequencies) in the eliciting of the various ERP components discussed above, although this aspect is not of direct relevance to this thesis. The background EEG activity is considered to reflect the brain's 'momentary' state of activity, and to subtly determine its response to the stimulus. Supporting evidence for this approach has been provided by numerous studies on the effect of stimulus presentation in particular brain states, on the morphology of ensuing ERP components. The studies to date have found that activity in the delta, theta and alpha EEG frequency ranges affects the major components of ERP, but activity in the beta range appears to have little direct involvement in their generation or morphology. Specifically, N1, N1P2 and P3 amplitudes, elicited by targets in auditory and visual paradigms, are directly related to the alpha amplitudes in the prestimulus period and inversely related to prestimulus delta and theta amplitudes. Furthermore, waves with latencies greater than 300 ms are directly related to prestimulus theta and delta activities.
It is now believed that the phase of EEG activity at stimulus onset affects the ERP outcomes. Studies have revealed that a regular pattern of stimulation can induce phase adjustments in the prestimulus EEG to a 'preferred' phase angle. In particular, it is suggested that within fixed ISI paradigms, the brain reorders the phase of its alpha activity (and perhaps its activity in other frequency ranges) so as to preferentially display negativity at stimulus onset, and this optimises its cortical excitability, and consequently affects behavioural and ERP outcomes. A more systematic investigation into the effects of phase synchronicity on ERP components is still in its developing stages.
____________________________________________________________________________________________________ |
