Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Review
, 95 (3), 773-80

Event-related Functional MRI: Past, Present, and Future

Affiliations
Review

Event-related Functional MRI: Past, Present, and Future

B R Rosen et al. Proc Natl Acad Sci U S A.

Abstract

The past two decades have seen an enormous growth in the field of human brain mapping. Investigators have extensively exploited techniques such as positron emission tomography and MRI to map patterns of brain activity based on changes in cerebral hemodynamics. However, until recently, most studies have investigated equilibrium changes in blood flow measured over time periods upward of 1 min. The advent of high-speed MRI methods, capable of imaging the entire brain with a temporal resolution of a few seconds, allows for brain mapping based on more transient aspects of the hemodynamic response. Today it is now possible to map changes in cerebrovascular parameters essentially in real time, conferring the ability to observe changes in brain state that occur over time periods of seconds. Furthermore, because robust hemodynamic alterations are detectable after neuronal stimuli lasting only a few tens of milliseconds, a new class of task paradigms designed to measure regional responses to single sensory or cognitive events can now be studied. Such "event related" functional MRI should provide for fundamentally new ways to interrogate brain function, and allow for the direct comparison and ultimately integration of data acquired by using more traditional behavioral and electrophysiological methods.

Figures

Figure 1
Figure 1
Adapted from Kwong et al. (20). BOLD contrast signal change is shown for a region of visual cortex during stimulation (on) and during rest (off). These data originally were used to demonstrate the application of BOLD contrast fMRI in normal human subjects. As can be seen, the rise time of the signal (indicated with arrows) is very rapid and has occurred after just a few seconds of stimulation, indicating that shorter stimulus events should be detectable.
Figure 2
Figure 2
Data from Robert Savoy and Kathleen O’Craven (25). BOLD contrast signal change are shown for visual stimuli of various brief durations. The three curves represent signal change for 34 msec, 100 msec, and 1,000 msec of stimuli, respectively. Importantly, clear signal change can be observed for events lasting as briefly as 34 msec.
Figure 3
Figure 3
Adapted from Dale and Buckner (31). (Upper) The raw BOLD fMRI signal evoked when either one, two, or three trials of visual checkerboard stimulation are presented. The trials were each 1 sec in duration and separated by 1 sec. The response increases and is prolonged with the addition of multiple trials, indicating it does not saturate going from one to three trials. (Lower) The explicit contribution of each individual trial by subtracting the one-trial condition from the two-trial condition (yielding the estimated response of the second trial) and the two-trial condition from the three-trial condition (yielding the estimated response of the third trial). The three estimated trials are roughly similar, although subtle but clear departures from linearity can be observed. This finding suggests the bold response can be shown to add linearly over trials, although the generalization of this finding to other brain regions and trial types is still an open question.
Figure 4
Figure 4
Examples of applications for event-related fMRI procedures are shown from several different laboratories. (A) Provided by Karl Friston (Wellcome Department of Neurology). A statistical activation map (horizontal section through the brain) of a subject hearing single auditory words is shown on the left with individual trial data from one activated voxel shown on the right (labeled adjusted data). The data are fit with a modeled hemodynamic response function (labeled fitted data), which shows a clear increase in relation to the trial onset. (B) Adapted from Clark et al. (43). Statistical activation maps are shown for three different stimuli types (scrambled faces, faces, and a target face that subjects are matching to a learned template). These stimuli were presented randomly in rapid succession, just a few seconds separating their onsets (see text). Methods based on statistical regression were used to separate the contributions of each stimulus type and generate statistical activation maps. (C) Adapted from Konishi et al. (40). An activation map showing a right prefrontal area active in no-go trials (no-go vs. go trials) of a response inhibition task. Pixels significantly activated 5 sec after no-go stimulus presentation are coded red. The area containing no-go dominant brain activity foci are enlarged, and the brain activity at several time points after the stimulus onset is shown for no-go trials (upper panels) and go trials (lower panels). (D) Adapted from Schacter et al. (49). Brain areas activated by averaged individual trials of a recognition task (see text) are shown based on two separate hemodynamic models with varied delays. Most areas are active at a relatively short delay of about 4–6 sec to peak. However, anterior prefrontal regions were active with a delay of about 8–9 sec.
Figure 5
Figure 5
Figure provided by Robert Savoy, Peter Bandettini, and Kathleen O’Craven (Massachusetts General Hospital). Activation within a region of visual cortex is shown for two separate conditions. In one condition (Left), the right visual hemifield stimulation proceeds the left by 500 msec. In the other condition (Middle), the left proceeds the right by 500 msec. Images from both of these conditions show signal changes with peaks spanning a considerable range in time (several seconds as indicated by scale on left). This intrinsic variance is too large to appreciate the 500-msec offset. However, once the data are normalized for this intrinsic variance by directly comparing the hemodynamic response from the two different lags within individual voxels, the offset between left and right hemifield stimulation can be appreciated (Right). Although not an explicit comparison across regions of brain, such a finding suggests that normalization of the hemodynamic lag within regions can allow small temporal offsets to be appreciated. These normalized offsets then can be compared across regions to make inferences about neuronal delay.
Figure 6
Figure 6
Snapshots of activity evoked by novel versus repeated words in a semantic judgment task at two different latencies are displayed on an “inflated” (54) left hemisphere. These data represent MEG data constrained by structural MRI and fMRI providing an image of cortical activity that is sensitive to fine spatial as well as temporal changes (46). By combining separate methods it is possible to obtain information unavailable from data produced by either method alone. Event-related fMRI, which can allow task paradigms that randomly intermix trial types at rapid rates, will further allow fMRI, MEG, and EEG to be combined in paradigms that are identical across methods.

Similar articles

See all similar articles

Cited by 63 PubMed Central articles

See all "Cited by" articles

Publication types

MeSH terms

LinkOut - more resources

Feedback