Speed of Processing in Sensory Systems

Measurement and Underlying Processes

The most obvious approach is to use some measurement of behavioral reaction time. This might involve determining how long it takes to name an object by using a voice key to record the moment at which the subject starts to vocalize. However, this type of responding takes quite a long time—at least 600 milliseconds (ms)—probably because accessing the verbal system is time consuming. There is also a large effect of word frequency on naming time—it is faster to name an “apple” than a “xylophone” for example—but it seems unlikely that this reflects a difference in sensory processing time.

An alternative approach would be to use a manual response, such as key pressing. For example, if a subject is asked to press a button each time a picture of an animal is flashed on a screen, minimal reaction times can be as short as 280 to 300 ms. But such values include not just sensory processing, but also the time needed to initiate and execute the motor response. How can researchers get a more direct measure of sensory processing speed?

One possibility is to look for brain signals in the electroencephalogram (EEG) or using magnetoen-cephalography (MEG) related to processing, but that precedes any behavioral responses. For example, in the animal detection task, there is a clear and robust difference in the event-related potentials (ERPs) generated on target and distractor trials starting about 150 ms after stimulus onset, implying that the underlying processing has been done by then. Although visible at frontal recording sites, it is now clear that the signals are actually generated in more posterior visual areas. Importantly, this differential response occurs even for images that the subject has never seen before and despite the fact that there is no way to predict what sort of animal will be shown.

Is this value of 150 ms a reasonable estimate of the speed of visual recognition? Intriguingly, recent experiments using eye movements instead of a hand response suggest that processing may be even faster. When two images are simultaneously flashed left and right of fixation, subjects can make reliable saccades (rapid eye movements) to the side where there is an animal in as little as 120 to 130 ms—substantially before the 150 ms differential EEG effect. If we allow at least 20 ms to initiate the eye movement, this implies that the presence of an animal in an image can be detected in no more than 100 ms. Such numbers put severe constraints on the underlying mechanisms.

...