| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Neuroradiology |
1 The Russell H. Morgan Department of Radiology and Radiological Sciences,
Division of Neuroradiology, Johns Hopkins Hospital, 600 N Wolfe St., Phipps
B-112, Baltimore, MD 21287.
2 Olin Neuropsychiatry Research Center, Institute of Living, 200 Retreat Ave.,
Hartford, CT 06106.
3 Yale University School of Medicine, 333 Cedar St., New Haven, CT 06510.
4 The Russell H. Morgan Department of Radiology and Radiological Sciences,
Division of Psychiatric Neuroimaging, Johns Hopkins Hospital, Baltimore, MD
21287.
Received June 19, 2003; accepted after revision May 10, 2004.
Address correspondence to D. M. Yousem.
OBJECTIVE. We sought to examine the correlation between reaction time and the amplitude of cortical activation during the performance of a visuomotor response-time task in a functional MRI (fMRI) experiment. We hypothesized that the fMRI blood oxygenation level-dependent (BOLD) amplitude may have a negative correlation with a subject's reaction time: the lower the amplitude within the cortical areas along the visuomotor pathway, the slower the response. A larger amplitude of the fMRI signal would reflect faster response times.
SUBJECTS AND METHODS. During a single-event fMRI experiment, the reaction times (in milliseconds) of 32 right-handed subjects responding to a visual cue were recorded. Analysis of the single-event paradigm using Statistical Parametric Mapping (SPM99) was performed, activation maps were produced for each subject, and then a random effects group analysis was performed. The maximum amplitudes of cortical activation (percent signal change) in four activated cortical regions were estimated and tabulated. The regions of interest included were the right and left occipital visual cortices, the supplementary motor area, and the left sensorimotor area. Simple and multiple regressions were performed between the mean reaction times of the subjects and the BOLD amplitudes in each region of interest and for the composite region of interest.
RESULTS. The results showed significant negative associations between the reaction times and maximum amplitudes in the right occipital, left occipital, and left sensorimotor area cortical regions (p < 0.05). However, no significant association was found between reaction times and the amplitude within the supplementary motor area. When the effects of age and sex on these associations were analyzed, we found that age had an impact on the results for individual regions of interest in the left occipital and left sensorimotor areas, but the composite amplitude of activation remained significantly correlated with reaction times.
CONCLUSION. The degree of signal change in BOLD fMRI response of the right occipital, left occipital, and left sensorimotor areas reflects the speed of performance during the visuomotor response time task by the subject. Thus, the amplitude of activation can be used as one parameter to assess change in function.
Several imaging techniques such as PET, MR spectroscopy, and functional MRI (fMRI) have been widely used to study the brain activity and its relation to human cognitive functions. PET has been used to examine changes in the regional cerebral blood flow as a consequence of motor and sensorimotor tasks [1, 2]. A recent study using MR spectroscopy found parallel changes in T2*-weighted image signal intensity and oxidative glucose metabolism during functional activation [3].
In addition, fMRI has been used to measure variations in the level of tissue oxygenation as a reflection of brain activation [4]. A recent study correlated the fMRI blood oxygenation level-dependent (BOLD) signal with the simultaneously measured neural activity through a microelectrode that recorded the stimulus-driven unit activity and the local field potential in anesthetized monkeys [5]. Researchers in that study found significant correlations between the fMRI BOLD magnitude and the neuronal and local field potential activity. Thus, the BOLD response reflected synaptic and postsynaptic phenomena and, in turn, the increased signal changes may reflect the activation of a greater number of synapses within the sensorimotor cortex [5].
Reaction time is the time between the onset of a stimulus and the motor reaction to that stimulus. A simple reaction time test includes the time from onset of the stimulus to the time of stimulus detection (perceptual latency) as well as the motor time, which is the time it takes to perform the motor task [6]. Many factors can affect the reaction time and the fMRI BOLD volume and signal change. Researchers have studied the volumes of cortical activation in fast- and slow-reacting subjects and the factors affecting the activation volume [7–9]. Oguz et al. [7] found that there was a greater activation volume in the motor and visual cortices in the fast-reaction-time group than in the slow-reaction-time group. However, to date, the degree of change in signal intensity during an fMRI experiment—the amplitude of the BOLD response—has received less attention.
We hypothesized that the fMRI BOLD amplitude (or percent signal change) may have a negative correlation with a subject's reaction time, such that longer reaction times are associated with lower percentages of activation.
Subjects and Methods
Subjects
Thirty-two healthy right-handed subjects (14 men and 18 women; age range,
23–85 years; mean age ± SD, 49.41 ± 17.78 years)
participated in this study. They were recruited from the healthy volunteer
registries of the Parkinson's Disease Research Center, an institutional
neuroradiology patient database, and advertisements in the print media. These
subjects had no metallic implants and no known neurologic or visual deficits.
Written informed consent from a protocol approved by the Johns Hopkins
Institutional Review Board was obtained from all subjects.
MRI Technique and Activation Tasks
Imaging was performed on a 1.5-T scanner (Gyroscan ACS-NT, Powertrak 6000,
Philips) equipped with 2.3-G/cm gradients and echo-planar imaging. A standard
head coil with foam padding to limit head motion was used. A screening
T2-weighted image (TR/TE, 4,000/102) was obtained in all patients to assess
for masses as well as for the presence and degree of white matter lesions.
Only subjects with no mass lesions and no significant white matter changes
(Cardiovascular Health Study grade 3 or less)
[10] were included in the
analysis (Appendix 1).
|
The fMRI protocol used a gradient-echo BOLD technique (1,000/39; flip angle, 90°; field of view, 24 cm; 360 time points in a 6-min scan). We acquired 12 scans angled parallel to the intercommissural line and included both visual as well as sensorimotor cortices with a 5-mm thickness and an interslice spacing of 1 mm using a matrix of 128 x 128. A TE of 39 msec does not supply enough time to fully sample the 128 x 128 k-space data matrix. Use of a 128 x 128 matrix requires partial (60%) acquisition of the k-space data and hence the actual spatial resolution exceeds the 1.875 x 1.875 mm pixel size within the slice.
The single-event paradigm, written in the E-prime (Psychology Software Tool) programming language, consisted of having a round circular multicolored visual cue randomly appear on the screen for 0.5 sec at either 20- or 30-sec intervals. Otherwise, a white fixation crosshair was constantly present in the center of a black background throughout the 6-min scanning period. The subjects were asked to fix their eyes on the crosshair and to tap a finger-press button with the index finger of their right hand as soon as they saw the visual cue. Reaction times during the experiment were measured (in milliseconds) from the button box and registered at the computer connected to the scanner.
Data Processing and Statistical Analysis
The single-event functional data processing was performed on Ultra
workstations (Sun Microsystems) using Statistical Parametric Mapping (SPM99)
(Wellcome Department of Cognitive Neurology of the Institute of Neurology)
implemented in Matlab (Mathworks)
[11,
12]. We used SPM99 software to
perform realignment for motion correction, normalization or deformation using
the standard brain template from the Montreal Neurological Institute and
conversion to the standard stereotaxic atlas of Talairach space
[13], smoothing at 5-mm
thickness, and performing data analysis (using an uncorrected threshold of
p < 0.001) for individual subjects. Random effects group analysis
was performed using the amplitude estimates (percent signal change) from the
32 subjects (a corrected threshold of p < 0.05). The images from
the fMRI data sets were displayed on standardized templates derived from
Montreal Neurological Institute and converted to the Talairach and Tournoux
atlas [13], after warping to
the atlas. These templates were used to uniformly display activation
localization on the individual fMRI maps and the group map (Figs.
1A and
1B).
|
|
Using a digital Talairach and Tournoux atlas template [14] labeled with Brodmann areas and gyri names, we defined the masks for the four regions of interest. After determining the region-of-interest masks for the left and right occipital visual cortices, supplementary motor area, and left sensorimotor area, the maximum amplitudes in each area using the group map were calculated. For the analysis of maximum amplitude, all voxels in the four regions of interest were analyzed even if they did not achieve statistical significance for activation in the SPM99 maps. However, by virtue of the emphasis on maximum amplitude, any voxels that exceeded the statistical threshold (p < 0.05) and were seen on the activation map were likely to have been included. A Bonferroni correction was not performed because the four areas of activation are not independent of each other, linked as they are by the physiologic response to the paradigm.
Results from the group analysis were tabulated. We used paired Student's t tests between the amplitudes of the four regions of interest as well as simple and multiple regression analyses of the mean reaction times with the amplitudes of the four regions of interest, accounting for age and sex of each subject. We then pooled the data for the regions of interest to obtain one composite value by averaging the maximum amplitudes of all four regions and applying a multiple regression analysis with amplitude as the dependent variable and composite reaction time, age, and sex as independent variables. We did the composite analysis because the four regions of interest are not independent of each other in their relationship in the activation pathway.
Results
The results showed the mean reaction time of the 32 subjects was 398.1 ± 73.7 msec. The average maximum amplitudes (percent signal change ± SEM) within the four areas in the group map are as follows: right occipital visual cortex, 0.55% ± 0.04%; left occipital visual cortex, 0.51% ± 0.04%; supplementary motor area, 0.23% ± 0.02%; and left sensorimotor area, 0.42% ± 0.04% (Fig. 2).
|
In comparing the maximum amplitudes of the four areas, we found significant differences between the supplementary motor area and the right occipital visual cortex, the left occipital visual cortex, and the left sensorimotor area (p < 0.0001); between the right occipital visual cortex and the left sensorimotor area (p < 0.001); and between the left occipital visual cortex and left sensorimotor area (p = 0.01). However, there was no significant difference between the maximum amplitudes in the right occipital and left occipital visual cortices (Table 1).
|
Simple regression analysis between the mean reaction times measured in milliseconds with the maximum amplitudes (percent signal change) showed significant negative associations with maximum right occipital, left occipital, and left sensorimotor amplitudes (p < 0.05); however, there was no significant association with supplementary motor area amplitude (Table 2 and Figs. 3A, 3B, 3C, and 3D). The association between the reaction time and the composite amplitude was also significant (p = 0.01) (Table 2 and Fig. 3E).
|
|
|
|
|
|
Using a multiple regression analysis to account for age and sex, we found that sex had no significant impact on these relationships and those correlations between right occipital, left occipital, and left sensorimotor amplitudes and reaction times remained significant. However, once age was introduced as an independent variable, only the right occipital amplitude and the reaction times remained statistically significant (p = 0.01), and there continued to be a relationship between left sensorimotor amplitude and reaction time, with a p value of 0.07. The composite amplitude still correlated significantly with reaction time (p = 0.037) for the combined independent variables of age and sex (Table 3).
|
Discussion
Studies of reaction time represent an important focus of interest because reaction times have a direct impact on the quality of life [15]. To understand the physiology of reaction times, we correlated the BOLD hemodynamic activity with reaction times. Many factors may influence reaction times, including the type, intensity, and background of the stimulus and the general health, age, sex, educational level or socioeconomic status, affective state, attentional–arousal state, caffeine usage, exercise level, blood glucose level, cardiovascular status, cardiovascular risk factors, and blood alcohol level of the subjects [16–22].
The Effect of Health Status on fMRI Results
A previous report noted that subjects with Parkinson's disease and slower
reaction times showed diminution in volume and amplitude of activation in the
left sensorimotor area compared with faster-reacting control subjects
[23]. On the other hand,
researchers who studied patients with Huntington's disease using visuospatial
task fMRI and FDG PET found that these patients had longer response times but
showed higher amplitudes than the healthy subjects. The researchers also found
that although these patients showed extensive parietal atrophy and reduced
resting glucose metabolism, they performed with an accuracy similar to that of
the controls but with a longer response times
[24]. The researchers
concluded that the increase in the percent signal change suggested that a
higher neuronal effort was required for patients with Huntington's disease to
reach a degree of accuracy similar to that of the control subjects; this
conclusion fitted well with the longer reaction time of the patients
[24]. In our study, our
subjects were healthy volunteers.
The Effect of Age on Reaction Time
D'Esposito et al. [9]
studied the effect of age on the neural activity with a visuomotor response
time task and found no significant difference between the younger and older
age groups in the magnitude of activation apparent in the activated voxels,
but they found that the number of suprathreshold voxels (volume of activation)
in the older subjects was four times less than the number in the younger
subjects. Those researchers also reported that the mean reaction time of the
older subjects who showed no activation had a trend toward being slower than
that of the older subjects who did show activation. Thus, they found that a
difference in age of the subjects may lead to significant difference in the
volume but not necessarily in the amplitude of fMRI BOLD activation
[9]. However, another study
showed that advanced age was associated with decreased amplitude of activation
[25]. These authors concluded
that the amplitude is related to the number of neurons involved in the
visuomotor task–related processing. Because reaction times often vary
with age, we used subjects of various ages in our study to have the greatest
range of reaction times to correlate with the percent signal change. However,
when age was as an independent variable, only the relationship between the
right occipital amplitude and reaction time remained statistically
significant. There continued to be a relationship between the left
sensorimotor amplitude and reaction time, with a p value of 0.07
When we pooled the data from the regions of interest together (because they are not independent of one another), we found that the composite amplitude correlated significantly with reaction time (p = 0.037) for the combined independent variables of age and sex.
The Effect of Sex on Reaction Time
The sex of the subjects can influence the amplitude of activation in fMRI
studies. Although researchers in one study using a visuomotor response time
task found no significant difference in volumes of activated voxels in
age-matched men and women [8],
earlier studies had shown that women have larger amplitudes of activation
[26,
27], whereas others had found
that the BOLD signal response was 38% lower in women during photic stimulation
paradigms [28]. Using a
multiple regression analysis, we found that sex had no significant impact on
these relationships and those correlations remained.
The Effect of Task and Paradigm on Reaction Time
The effects of alternating and continuous single-finger opposition task
designs (rest and activation epochs) on BOLD signal contrast have been
reported. Mohamed et al. [29]
found that signal intensity was more robust in alternating patterns; however,
these authors did not correlate their findings with the reaction time. In our
study, we used a fixed stimulus for all subjects and measured the response or
output in terms of time rather than the input in terms of stimulus intensity
or pattern.
We chose a visuomotor reaction time task for its simplicity and well-described place in the literature to date. Also, this task is well understood, and the fMRI findings can be compared with those of previously performed behavioral studies [30–34]. We also preferred the event-related paradigm so as to reduce the effect of anticipation or accommodation that might produce an increase in neural activity within the motor cortex and supplementary motor area before the onset of stimulus [35, 36].
Analysis of Our Results
In this study, we focused on measuring only the amplitude (or the
percentage) of fMRI BOLD signal change rather than the volume of the BOLD
signal. Investigators in a previous study correlated the reaction time with
the volumes of activation in the visual and sensorimotor cortices in fast- and
slow-reacting groups [7]. Those
authors found that the volumes of activation were higher in the fast-reacting
subjects than the slow-reacting ones. However, we performed a regression
analysis between reaction time and the amplitudes of the regions of interest
without grouping subjects into fast- and slow-reacting subjects. This type of
analysis is more robust than separating subjects into two groups and
performing a simple Student's t test.
On calculating the percent signal change within an area in the group map, we identified significant differences in the maximum BOLD fMRI signal within the four regions of interest studied. These differences in the maximum amplitude were more evident between supplementary motor area (with low BOLD contrast) and the other three regions studied. Although there was no significant association between the BOLD amplitude in supplementary motor area and the reaction time, the t value was negative, with a negative slope on regression analysis, indicating that the correlation was in the same direction (i.e., the slower the response, the lower the BOLD amplitude).
Furthermore, one study used PET to investigate brain activation from stimuli using different sensory techniques to determine which neural populations are responsible for the speed of reaction. In agreement with our results, they found significant negative correlation of the regional cerebral blood flow with mean reaction times (i.e., the shorter the reaction times, the greater the change in blood flow, analogous to our BOLD amplitude findings) [37]. Comparing our data for the four regions of interest, we found that the association between reaction time and amplitude was significant in the visual and motor cortices (p < 0.05).
Other researchers who used electromyography to study the relationship of different reaction tasks have shown that increased excitability of the corticospinal output is not required to speed up reaction times and that the time taken to discharge cortical output elements is relatively unimportant compared with the time needed to process the sensory (e.g., visual) input and link it to the motor output [38]. In our study, electromyography was not performed; however; its use should be considered in future studies to exclude the peripheral motor output function as a factor that might affect the reaction times or speed of performance. Also, further studies on the time course and latencies between the sensory and motor cortices should be studied and correlated with amplitude as well as reaction time
Why study amplitudes of activation and not volumes of activation? Volumes of activation have shown a variable response to reaction times [9]. Does brain activation decrease with shorter response times because of more efficient use of brain pathways, or does an increase in brain activation occur because more neurons are recruited for the successful performance of the task? These different theories have led to difficulties in understanding how to analyze and interpret fMRI results in a subject population at risk for cognitive decline or propensity for neurodegenerative disorders. Will volumes of activation go up or down? If amplitude of activation can be used instead as a more reliable means for determining the effectiveness of the brain in performing a task that is a predictor for health and disease, it would supplant the volumetric approach to fMRI. Our study is at the forefront of this approach.
Whether the differences in the BOLD fMRI amplitude estimates are due to differences in the cerebral hemodynamics such as regional cerebral blood flow or due to the factors involved in neurovascular coupling, our study has shown a negative association between the right occipital, left occipital, and left sensorimotor amplitudes of the fMRI activation and a subject's reaction time, thus reflecting the subject's performance. The sex of the patient does not appear to influence this relationship, but age does. However, when the entire circuit is viewed as a whole with a composite amplitude of activation for the four regions of interest, reaction times and amplitudes of activation continue to have a significant correlation. The amplitude of activation may be of value as one parameter with which to assess reaction time performance before and after experimental interventions.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
