Academic journal article
By Melnikov, Alexander V.; Singer, Robert N.
Research Quarterly for Exercise and Sport , Vol. 69, No. 4
Recent approaches to research in the neurosciences and cognitive psychology have led to the refinement of chronometric studies of human information processing. An important approach for studying mental chronometry, especially in the framework of information-processing models, is event-related potential (ERP). In general, ERP is defined as "scalp recorded electrical potentials that typically originate as postsynaptic potentials in populations of cortical neurons that are activated during sensory, motor, or cognitive processing" (Luck & Hillyard, 1994, p. 1001). Event-related potentials reveal not only the "timing" of a particular cognitive process relative to a specified event but also its type and degree of activation. Although ERP data provide important insights about behavior, applications have generally been limited, thus far.
A primary reason for this is a rather strict limitation on the measurement procedure. There are requirements for numerous repetitions during data collection, which are used later to average the signals. The procedure of averaging demands precise synchronization of the electroencephalographic (EEG) signal with the presented stimulus, because changes in the EEG associated with stimulus are relatively small and usually obscured by much larger spontaneous brain waves and rhythms. Summarizing the EEG epochs and averaging them over numerous presentations of the same stimulus enhances EEG components caused by the stimulus and reduces random and spontaneous waves. Furthermore, the nature of the EEG recording requires minimal interference from eye movements and eye blinks that might distort the data. These stringent requirements make it hard to use the method in dynamic situations, in which an action sequence may contain various target stimuli and potential responses compared to restricted and artificial laboratory settings in which simple computer-generated visual stimuli are tightly bound to one or two responses. Furthermore, applying the results obtained in the laboratory to real-world situations could be controversial, unless there is reasonable congruence with what is tested in the laboratory and what is demanded in the actual performance setting.
In general, there are many reasons why ERP studies have appeared infrequently in applied research, and why questions about ecological validity arise every time the results of laboratory experiments with ERP are evaluated for implications to real-life dynamic situations. In this article, an alternative methodology for studying ERP is proposed. It allows investigators to measure ERP in situations closer to actuality, making ERP analysis more useful.
Defining and Measuring ERP Components
ERPs are described in terms of polarities and latencies of components in the wave form. That is why every component is typically named by a code indicating its polarity and latency. For instance, P3, or P300, is a third major positive component which takes place approximately 300 ms after stimulus onset. It is important that every ERP component is sensitive to the different cognitive processes involved in information processing. The first negative component, N1, after stimulus onset, which usually has a latency of 100 ms, mainly corresponds to selective attention. The second positive component, P2, with a latency around 200 ms, usually reflects an early processing phase or low-level attentional demands (Polich, 1993; Raney, 1993). The amplitudes of these components, and also differences between them (N1 P2), are used for evaluating the intensity of the cognitive load while executing a task. For example, in experiments in which the task was for participants to remember as much as possible of written passages for later recall, an increase of the N1 - P2 amplitude was interpreted as indicating lesser cognitive load (Raney, 1993).
The second negative component, N2, usually takes place approximately 200 ms after onset of the stimulus and has been associated with pattern recognition and stimulus classification (Simpson, Vaughn, & Macht, 1982) or with filtering spatial information. …