|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
NEGATIVE BOLD AND AGING: AN fMRI STUDY
KEITH MATTHEW MCGREGOR
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
Keith Matthew McGregor
This thesis is dedicated to Kristi Michelle Stahnke.
This project could not have been completed without the guidance and friendship of
Dr. Keith White, to whom I am so very much grateful. I would also like to thank Dr.
Bruce Crosson for his steadfast leadership and generosity and for providing the nurturing
environment for this endeavor. Finally, I would like to acknowledge Dr. Ira Fischler, Dr.
Keith Berg, Michelle Benjamin, Dr. Jason Craggs, Dr. Timothy Conway, my family and
friends for their support and assistance throughout the completion of this project.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iv
LIST OF TABLES .............. .................. ............ .......... ................. vii
L IST O F FIG U R E S .............. ............................ ............. ........... ... ........ viii
ABSTRACT ........ .............. ............. ...... .......... .......... ix
1 IN T R O D U C T IO N ............................................................................. .....................
Functional M agnetic Resonance Im aging..................................... ...............
Negative BOLD and the Present Study ....................................... ............... 1
R ole of Interhem ispheric Connectivity ................................................. ...............3
H y p oth eses .................................................................................. . 4
2 M E TH O D S .................................................................5
P articip ants .................................................................................. . 5
Procedure ........................................ .. ....... ......... 6
Event-Related Block (ER-Block) Imaging Paradigm ......................................6
A pparatus ...................... ...................................................................... ........ 10
Functional Imaging Data Analysis ............................................11
Hemodynamic response function analysis........................... ............... 12
3 R E SU L T S ...........................................................................................................13
B ehavioral P perform ance................................................................................... 13
Im aging A naly sis .............. ..... ..............................................13
Hemodynamic Response Function Analysis ...................................... 15
M ultivariate AN OVA ...................... ................................... .... ...........18
4 D ISC U SSIO N ...................................................... 19
L im itation s .................. ............................................................ 2 1
Implications and Future Directions.............................. ...... ........ 22
L IST O F R E F E R E N C E S ...................................... .................................... ....................23
B IO G R A PH IC A L SK E TCH ...................................................................... ..................29
LIST OF TABLES
1 Participant characteristics ......................................................... ...............
2 Talairach coordinates and volumes of active voxel clusters in the right Ml
during learned response executions................... .... ... .. ............... 16
LIST OF FIGURES
1 Sample stimuli for response conditions are shown for: (A) GO; (B) NO GO; and
(C) N O V EL .......................................... .............................. 8
2 Estimated hemodynamic response functions in a younger participant (s12)...........14
3 Estimated hemodynamic response functions in an older p0061rticipant (s01)........15
4 These four graphs show representations of the estimated hemodynamic response
function av erages.......... ................................................................... ....... .. .... 17
5 Group means overlayed on axial images....................................... ............... 18
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
NEGATIVE BOLD AND AGING: AN fMRI INVESTIGATION
Keith Matthew McGregor
Chair: Keith D. White
Major Department: Psychology
The goal of this event-related fMRI study was to investigate age differences using a
task previously reported to elicit negative BOLD signals. Six (to date) right-handed
individuals from two age groups were trained to perform a right hand, externally paced,
12 finger movements button press sequence. This task was performed during each of
three response conditions: (a) immediate execution of the learned sequence (GO); (b)
suppression of the learned sequence (NoGO); (c) initial suppression of the learned
sequence followed by execution (DelayGO). Four regions of interest (ROI) were selected
for analysis: both primary motor areas (LM1, RM1), and both supplementary motor areas
(LSMA, RSMA). Deconvolution analysis was used to estimate hemodynamic response
profiles (HDRs) of maximally active voxels in each ROI for each response condition.
Repeated measures analysis of variance of HDR profiles was performed across age
groups, response conditions, and ROI. Younger subjects exhibited positive BOLD signals
in LM1, LSMA and RSMA with negative BOLD signals in LM1 during GO and
DelayGO conditions, as expected. Older adults, in contrast, showed positive BOLD
signals in all four ROIs during those conditions. Amplitudes and times to peak of the
HDRs did not differ between age groups. These preliminary results indicate age
differences in activation of R Ml cortex during execution of the right-hand learned
button press sequence. These differences in BOLD signal across age could be related to
the hemispheric asymmetry reduction in older adults (HAROLD) model.
Functional Magnetic Resonance Imaging
Functional magnetic resonance imaging (fMRI) has greatly enhanced our
understanding of brain activation patterns during cognitive tasks. Blood oxygen level-
dependent (BOLD) contrast has been the leading technique used by researchers to
investigate neural correlates of cognitive activity using fMRI. Synaptic activity, whether
excitatory or inhibitory, causes an increase in the local concentration of deoxygenated
hemoglobin (deoxyhemoglobin) due to accelerated oxygen metabolism.
Deoxyhemoglobin reduces the bulk magnetization of nearby protons, reducing the
measured MR signal through a dephasing effect. An overcompensation of oxygenated
blood is then sent to the metabolically active area in a hemodynamic response (HDR)
coupled to the increased oxygen demand. Local deoxyhemoglobin is diluted, raising the
MR signal above baseline. As the oxygen demand returns to baseline levels, so too will
the MR signal gradually return to its baseline. The HDR results in a roughly bell-shaped
MR signal change, a positive BOLD response (PBR) over the course of 10 15 seconds
for a relatively brief interval of synaptic activity. A measurement volume (voxel)
exhibiting PBR signals during tasks is reported as being activated (Ogawa et al. 1990;
Logothetis et al. 2001).
Negative BOLD and the Present Study
A negative BOLD response (NBR) appears like a "mirror image" of the typical
PBR described above, that is a U-shaped rather than bell-shaped time course of MR
signal change. NBR indicates a decrease in metabolic/neuronal activity in the area in
which it is found (Shmuel et al. 2002; Stefanovic et al. 2004; Amedi, Malach, & Pascual-
Leone, 2005). Specifically, Stefanovic et al. (2004) correlated NBR to decreases of
regional metabolic rate of oxygen consumption (CMRO2). In some studies, voxels
exhibiting NBR are included as being activated because they pass a statistical test for
differing reliably from baseline.
Older adults have higher prevalence of NBR as compared to younger adults during
both motor and visual tasks (Aizenstein et al. 2005). Aizenstein et al. (2004) contended
that spatial averaging over a mixture of voxels having different directions of activations
(i.e., NBR versus PBR) could possibly account for previous findings of decreasing
amplitude of BOLD response with increasing age in the visual (Huettel et al. 2001; Ross
et al. 1997; Buckner et al. 2000) and motor cortices (D'Esposito et al. 1999; Hesselman et
al. 2001; Buckner et al. 2000). Older adults have higher inter-individual variability in
BOLD contrast, consistent with individual variations in the mixture of NBR and PBR
(see D'Esposito et al. 2003 for review). However, we are unaware of any studies that
were specifically designed to elicit NBR for the purpose of studying age effects. Many
studies have found NBRs serendipitously rather than by design (Allison et al. 2000;
Roether et al. 2002; Born et al. 2002; Aizenstein et al. 2004; Amedi, Malach, & Pascual-
Leone, 2005) while studies designed to elicit reliable NBRs have not included older
adults (Shmuel et al. 2002; Hamzei et al. 2002; Hamzei et al. 2005; Stefanovic et al.
2004; Hummel et al. 2004; Newton et al. 2005). We have adapted methods from the
latter studies, making some modifications to accommodate older adults, for the present
The NBRs are often elicited in a unimanual motor task wherein the contralateral
(hand controlling) hemisphere's primary motor cortex shows PBR, as expected, but at the
same time and in the homologous location, the ipsilateral Ml shows NBR (Hamzei et al.
2002; Stefanovic et al. 2004; Hamzei et al. in press; Gardner et al. 2005). The amplitude
of the NBR has been shown to be reasonably proportional to, although comparatively
smaller than, than the other hemisphere's PBR (Newton et al. 2005).
Role of Interhemispheric Connectivity
Interhemispheric connectivity is widely believed to be the source of the NBR
(Allison et al. 2000; Shmuel et al. 2002; Stefanovic, et al, 2004). It is well known that
transcallosal synapses are mainly glutamatergic and exhibit excitatory effects on their
target (Gerloff et al. 1998). Thus, cortical cell bodies resident in one hemisphere, which
might extend actively firing axons transcallosally, generate increased metabolic activity
at the target location due to the release of excitatory neurotransmitter. At least one more
level of connections must exist involving inhibitory neurotransmission in order to reduce
neural/metabolic activity. Because cortical inhibitory intemeurons generally synapse
locally (nearly all axons extend <200 microns), the accompanying increase in synaptic
activity (due to both the transcallosal excitatory inputs and the local inhibitory outputs)
will necessarily require a subsequent increase in blood flow resulting in PBR in the target
vicinity. One possibility is that these inter-hemispheric excitatory axons terminate on
dendrites of inhibitory intemeurons that in turn project to multiple pyramidal cells
(Beaulieu and Colonnier, 1985) via axo-somatic or axo-axonal terminal branches (Koos
and Tepper, 1999). Therefore, the downstream synaptic outputs of the pyramidal cells
must be presumed to be disproportionately reduced in order to outweigh the PBR evoked
by the synaptic activities bringing about the volume-averaged synaptic activity reduction.
Negative BOLD could also be exhibited one level downstream of inhibitory
synapses distant from the target vicinity, for example in the thalamus, where reduced
spike frequency would result in reduction of those thalamic axons' synaptic outputs at the
cortex. This alternative means, possibly involving cortico-striatal-thalamo-cortical loops
(Alexander et al. 1990), would also reduce cortical metabolic activity by reducing its
synaptic inputs. If this alternative were to have merit then not only primary motor (Ml)
but also supplementary motor (SMA) cortices would be likely to evidence activations in
at least one but probably both hemispheres. Therefore, regions of interest for the present
study will include both hemispheres' Ml and SMA.
Furthermore, it is well known that aging has a large influence on the striatum,
specifically that dopamine uptake in caudate and putamen declines with age as indexed
by levels of dopamine transporters (Kish et al. 1992; van Dyck et al. 1995; van Dyck et
al. 2002; Erixon-Lindroth et al. 2005). Reduction of dopaminergic input onto the Dl
receptors in the "direct" path (striatum, globus pallidus internal, thalamus) would be
expected to cause reduced net inhibition of the thalamus, and thus increased cortical
input. This leads us to hypothesize that older adults will have a reduced magnitude of
Data from 12 right-handed participants were used to estimate the presence of
negative BOLD in two age groups (6 Young, mean age = 22 yrs, sd = 3.09 yrs; 6 Old,
mean age = 71 yrs, sd = 6.49 yrs), see Table 1. All participants were prescreened by self-
report for claustrophobia, contraindicated medications or metal implants, pregnancy
(when applicable), and pre-existing neurological or psychiatric disorders. Participants
with over two years of recent (within five years) piano playing experience were also
excluded due to potential variation of signal dynamics (Ridding et al. 2000; Krampe,
2002). All participants were recruited and provided written consent after the experimental
procedures were fully explained to them in accordance with the Internal Review Board of
the University of Florida. Cognitive functioning in the older adults was assessed with the
Mini Mental Status Exam (MMSE) with a minimum score of 26 needed for participation.
All participants were asked to avoid caffeine for at least six hours prior to scanning (see
Laurienti et al. 2003; see also Laurienti et al. 2004).
Table 1: Participant characteristics.
N 6 (4 female) 6 (3 female)
Age Range: 19-27 60-76
SD: 3.09 6.49
Mean: 24.0 71.0
MMSE mean (range) N/A 29.2 (28-30)
Prior to scanning, participants were trained to "overlearn" a 12 movement button
sequence with the fingers of their right hand. Task training outside the scanner room
began with the participants learning the proper numerical representation for each finger
of the right hand. The thumb was designated as 1 with integer increments for the
subsequent digits (e .g index finger 2, middle finger -3, etc.). The target sequence was
as follows: 5-4-2-3-4-4-2-5-2-3-3-4. Sequenced button press training was performed on a
five-button Model 300 PST Response Box (Psychology Software Tools, Pittsburgh, PA)
and consisted of three phases. The first, a "familiarity phase", required participants to
self-pace sequence practice while viewing a display of the sequence on a laptop
computer. This was followed by a "memorization phase" during which participants
continued self-paced sequence presses, but now with the visual sequence display
removed. Finally, in a cuedd phase", participants were asked to press button sequences in
rhythm with the 1 Hz flashes of a star displayed on a computer screen. After completing
target sequence executions 10 times in a row without error, participants advanced to the
next progressive training stage. Scanning commenced after the successful completion of
the cued phase. All participants were able to train to criterion within the allotted 1-hour
Event-Related Block (ER-Block) Imaging Paradigm
In the scanner, participants lay supine viewing a fixation cross in the center of the
screen at which they are instructed to maintain fixation throughout scanning. During
each run, button press trials begin after a pseudo-randomly generated resting-state
interval of 6, 7 or 8 TR (10.2, 11.9, OR 13.6 sec). Trials are consisted of four types: GO,
NO GO, DELAY, and NOVEL. The GO, NO GO, and DELAY trials are together
referred to as learned sequence trials, as they involve the execution or suppression of the
target finger movements. The NOVEL trial was added to mitigate boredom during
scanning (see Hummel et al. 2004). Subjects are informed of the trial type through two
different "prepare" cues. Participants are notified of an impending learned sequence trial
by the presentation of a single green "prepare" star in the top center of the screen. After
1 TR (1.7 seconds), the prepare star disappears and one of three trial types commences:
GO, NO GO, or DELAY. Participants were cued of the NOVEL sequence in a similar
manner except that a yellow star is used instead of a green star. An explanation of each
trial type follows and sample images are presented in Figure 1.
In a GO trial, a green "go" star begins to flash at 1 Hz in the center of the screen at
the end of the "prepare" period. The participant was instructed to press the learned
button sequence by timing their button presses with 12 flashes of the green go star. Stars
remain on the screen for 500 ms. We instructed participants to continue through the
sequence even if they believed they made an error at any point during its execution.
In a NO GO trial, a red, "no go" star is displayed in the screen's center cueing the
participant to refrain from pressing the button press sequence.
In a DELAY, as in a NO GO trial, a red "no go" star follows the aforementioned
"prepare" cue. However, after a delay of either 1 or 2 TR (1.7 or 3.4 sec), the red star is
replaced by 12 flashes (also at 1 Hz) of a green "go" star, instructing the participant to
begin pressing the sequence in rhythm with the flashing star. We instructed participants
to maintain attention to the red star as "it may change to green at any point during its
presentation". After the trial's prepare period, the NOVEL response trial requires
participants to press pseudorandomly generated integers ranging from 2-5. The integers
are individually flashed near the bottom of the screen, each remaining visible for 500 ms.
12 of these random integers are flashed during the trial at 1 Hz. The time of display for
each integer is identical to the "go" star flashes in the learned response executions
Figure 1: Sample stimuli for response conditions are shown for: (A) GO; (B) NO GO;
and (C) NOVEL.
In the scanner room, participants were reminded of the task procedures prior to
entering the scanner apparatus and then again immediately before commencement of
functional imaging. After structural acquisition but prior to EPI image acquisition,
participants performed a practice run with representative stimuli. The scanner operator
visually monitored participant performance on an LCD computer monitor in the control
room which displayed the button responses. After successful completion of a full trial
run (6 minutes 41 seconds), functional imaging proceeded. No participants required
additional training or reported discomfort with the task as verified by verbal report
immediately prior to EPI acquisition.
Each of the 6 functional scanning runs (6 min 41 sec each; 1368 total images)
included 12 stimulus presentations in pseudo-randomized order for a total of 72 trials in
the imaging session. Of these presentations, 20 trials each were GO, NO GO, and
DELAY. DELAY trials were further divided into 10 each of 1-TR and 2-TR delays. As
they were mainly used in the paradigm to mitigate boredom, NOVEL trials were
presented only 12 times (2 per run). Pseudo-randomization of the NOVEL button presses
was done during paradigm design and, as such, all subjects received the same NOVEL
Older and younger adults were not expected to vary on behavioral performance
(accuracy and reaction time) due the training procedures and the timed nature of the task
(Janahashi et al. 1995; Rao et al. 1993). Button press performance during scanning was
evaluated on an individual response basis with each button press in the sequence taken as
a response point (52 trials X 12 presses/trial = 624 total response points per participant).
Evaluation of accuracy of each response point takes into consideration not only the
specific button pressed, but also the reaction time. If a participant did not make a
response within 500ms of the cue or made a response within 300ms prior to the cue, then
the button press is marked as incorrect regardless of key identity. Evaluation of errors
was completed with E-Prime 1.1, Microsoft Excel, and simple Perl 5.0 scripts written by
the author. Error rates for tasks of the current study's types have been reported to be less
than 10 percent (Liepert et al. 2001; Verstynen et al. 2005; Gerloff et al. 1998; Hummel
et al. 2004). Reaction times and error rates were averaged across age groups for each
condition and independent samples t-tests were used to compare these means.
All scanning was performed on a Siemens Allegra 3T MRI machine located at the
McKnight Brain Institute of the University of Florida. After each participant was fitted
with an MR-compatible headset (Resonance Technologies Inc.), their head was
positioned within a standard quadrature RF head coil with padding to limit head motion.
First surface mirrors attached to the head coil were positioned above the participant's
eyes and rotated to allow viewing along the scanner bore axis. After the participant had
been moved into the bore, three single slice orthogonally oriented scout images were
obtained to ensure that all parts of the head were within the homogeneous portion of the
magnetic field. The scout images (TR = 18, TE = 4.28, FA = 25 degrees, 256x256,
multiplane slice thickness = 3.0 mm, FOV = 280 mm) were also used to prescribe slice
alignment parallel to the participant's interhemispheric longitudinal fissure. Next, a
magnetic resonance time-of-flight angiogram (MRA) was acquired using TR = 23, TE =
6.15, FA = 30 degrees, matrix = 256x256, axial slice thickness = 5.0 mm, FOV = 240
mm. The MRA allowed slice prescriptions aligned parallel to the anterior commissure -
posterior commissure line as well as to the longitudinal fissure. Slice prescriptions for all
other image acquisitions were copied from the MRA. High resolution anatomic images
were next obtained using a Siemens MPRAGE sequence with TR = 23 ms, TE = 6 ms,
flip angle = 25 degrees, matrix = 256x192, 128 axial slices at 1.3 mm thickness (no gap),
FOV = 240 mm. Lastly, the behavioral task was executed during multiple acquisitions of
gradient echo echo-planar images (EPI) with the following parameters: TR = 1700 ms,
TE = 50 ms, flip angle = 70 degrees, matrix = 64x64, 32 axial slices at 4 mm thickness,
FOV = 240 mm.
Participants viewed stimuli on an IFIS-SA presentation system (InVivo Systems,
Gainesville, FL) using a Dell model 2100MP data projector, rear projection screen, and
first surface mirror display system. Stimuli were projected at 1024 x 768 pixel
resolution, visible to the participant through two mirrors (one on the head coil, another
outside the scanner bore). A MR-compatible button response pad, part of the IFIS-SA
system, attached to the participant's hand and wrist was used to record individual finger
presses throughout the duration of scanning.
Functional Imaging Data Analysis
Imaging analysis was performed with AFNI software (Cox, 1996). The eight initial
magnetization equilibration images (called discarded acquisitions or "disdaqs") from
each functional run were omitted from subsequent analysis. Using a 3-dimensional rigid
body registration, functional images were aligned to a base image from the functional
volume acquired closest in time to structural images. Linear trends in the time series of
each run were removed, and all imaging runs were then concatenated. Deconvolution, a
multiple regression analysis procedure for estimation of hemodynamic response in each
voxel, was performed using AFNI's 3dDeconvolve program. As the GO and DELAY
conditions were very much alike in their execution, the two conditions were combined
into a "learned response execution" condition to enhance statistical power (n = 40
response trials). (A previous analysis of activation patterns both within and across age
groups confirmed that GO and DELAY response conditions do not significantly vary in
correlates of functional activity.) As negative BOLD is believed to be a "mirror image"
of the positive BOLD signal, the non-directional F-statistic was chosen as the output
statistical measure. Activation for a given voxel was determined if the F statistic from the
deconvolution exceeded a value of 3.5 (p<.00001, uncorrected for multiple comparisons).
Hemodynamic response function analysis
We hypothesized that younger adults show a different NBR in R Ml than older
adults during learned response execution. To test this, we compared the time courses of
the estimated hemodynamic response functions as calculated by the 3dDeconvolve
application (Ward, 2002) across older and younger adults. These hemodynamic response
function (HRF) estimates were created as follows. Four regions of interest were
identified: left and right primary motor cortex (L/R Ml), and left and right supplementary
motor cortex (L/R SMA). For each participant and each region of interest, the voxel with
the highest F-statistic was identified and the 16 numerical values from its estimated HRF
time course were saved. An additional four active voxels (F > 3.5, p<.00001) in spatial
contiguity with the maximally active voxel were also identified and 16 numerical values
from each voxel's estimated HRF time courses were also saved. The five saved voxel-
wise estimates of the HRF were averaged for each of the 16 time points, for each
participant and region of interest, to represent the time course of the estimated
hemodynamic response under those provisos. This HRF averaging procedure was carried
out for all 4 regions of interest in all 12 participants. A repeated measures ANOVA (2 X
4 X 16) compared age (young, old); region of interest (L Ml, R Ml, L SMA, R SMA),
and time (16 HRF points).
While learning the button-press response sequence prior to scanning, the older
participants made more errors (older mean = 90.5; younger mean = 32.5, t(10) = 2.38,
p<.04) and required longer to train to criterion (older mean = 34.2 minutes, younger mean
= 24.6 minutes; t(10) = 2.61, p<.03). During scanning, however, there were no
significant differences due to age in either number of errors or reaction times on GO, NO
GO, and DELAY trials. Older participants had slightly longer reaction times on the
NOVEL trials (older mean = 405 msec, younger mean = 365 msec; t(10) = 4.66, p<.01).
The generally comparable accuracies and speeds of executing the learned response
sequence are noteworthy because behavioral performance does not require inclusion as a
covariate in subsequent imaging analyses.
During processing of the fMRI images for individual participants, it became readily
apparent that negative BOLD responses existed, as hypothesized, in the right (ipsilateral)
motor cortex (R Ml) of the younger subjects (see an example in Figure 2) but positive
BOLD response instead existed in the R Ml of the older participants (see an example in
Figure 3). Figures 2 and 3 each present for individual subjects (younger, s12; older, s01):
(a) five examples of estimated hemodynamic response functions (HRFs) from voxels in R
Ml, and (b) sample fMRI axial slices incorporating the "hand knob" region of R Ml. All
brain images are in neurological convention (left = left on axial view). Voxels within R
Ml were selected to form a contiguity group using the following criteria: First, the voxel
having the highest F-statistic associated with deconvolution relative to GO and DELAY
(response execution) conditions combined was identified as the "max-voxel". Next, four
acquisition voxels touching max-voxel and having F > 3.5 were identified. The five
HRFs shown in Figure 2(B) and Figure 3(B) were obtained from the R Ml contiguity
groups for s12 (younger) and s01 (older), respectively.
Figure 2: Estimated hemodynamic response functions in a younger participant (s12) (A).
Five estimated hemodynamic response functions from four voxel contiguous
to the most highly active voxel (bottom HRF in A) in R Ml of s12 (young
adult) during learned response execution conditions. Voxel locations in
cortex are represented in (B) with colors coding probability associated with
the F statistics (yellow: p<.0001; orange: p<.001; red: p<.005) from
deconvolution analysis.Contiguity group multivariate ANOVA
Figure 3: Estimated hemodynamic response functions in an older participant (s01) (A)
Five estimated hemodynamic response functions from four voxel contiguous
to the most highly active voxel (bottom HRF in A) in R Ml of s01 (older
adult) during learned response execution conditions. Voxel locations in
cortex are represented in (B) with colors coding probability associated with
the F statistics (yellow: <.0001; orange: p<.001; red: p<.005) from
Hemodynamic Response Function Analysis
The example HRFs in Figure 2(B) each resemble negative BOLD responses, while
the example HRFs in Figure 3(B) are typical (positive) BOLD responses. This striking
age-related difference in the polarity of BOLD hemodynamic response in these
individuals has been confirmed more generally by: (1) localizing the most statistically
robust contiguity groups, as described above, within four hypothesis-driven regions of
interest (specifically the left and right Ml and the left and right supplementary motor
cortices (SMA), and (2) extracting HRFs from each voxel of these contiguity groups,
calculating area-under-the-curve for each HRF and subjecting these to multi-factor
Table 2 presents localizations and activation volumes of R Ml contiguity groups,
as defined above, for all subjects for deconvolutions relative to response execution
conditions. Localizations are given in Talairach coordinates. Activation volumes are
based on 1.8 mm contiguity radius for 1 mm3 isotropic voxels centered on max-voxel.
All six younger adults evidenced negative BOLD in the R Ml, while all but one older
adult evidenced positive BOLD in the same area.
Table 2: Talairach coordinates and volumes of active voxel clusters in the right Ml
during learned response executions. Clusters were defined using an F-statistic
threshold set at 3.5 (p<.00001, uncorrected) with a contiguity radius of 1.8
mm3. and volume threshold of 50 [tL. Negative BOLD clusters are shown in
Older Talairach Volume Younger Talairach Volume
Adults Coordinates Adults Coordinates
S01 (-46, -2, 53) 285 LtL s07 (-46 ,-2, 53) 64 pL
S02 (-35, 7, 56) 106 [L s08 (-58, 2, 42) 83 pL
S03 (-30, 14, 52) 110 lL s09 (-39, 19, 40) 74 pL
S04 (-28, 26,57) 454 [tL sl0 (-43, 11,50) 109 pL
S05 (-23, 6, 58) 118 pL sll (-23, 6, 58) 92 L
S06 (-32, 26, 52) 325 tL s12 (-35, 36, 58) 193 pL
Four regions of interest (ROI), left and right Ml and left and right SMA, were
selected within each subject and five-voxel contiguity groups were identified for each
ROI. HRFs were obtained from each of these 20 voxels per subject. Figure 4 shows the
averages of these HRF time courses within each ROI and age group. The ordinates on
Figure 4 show magnitude of HRFs from deconvolution relative to response execution
events. Each ordinate is plotted against image acquisition time (in TRs) relative to the
response execution event, and compares younger and older age groups. This kind of plot
is shown for each ROI. Starting from the top left panel, representing L Ml, large positive
BOLD HRFs were found as expected for both younger and older groups..
Figure 4: These four graphs show representations of the estimated hemodynamic
response function averages of the most highly active voxel and four of its
suprathreshold contiguous neighbors for older and younger adults during
learned response execution. Each of the four graphs shows a different region
of interest (clockwise from top left: L Ml, LSMA, RSMA, RM1). The
ordinate values are the magnitude of the hemodynamic response function
(HRF) shown against time (in TR units) relative to the response event. Large
positive BOLD responses can be seen for older and younger adults in L Ml.
Moderate positive BOLD responses are shown for each age group for SMA
bilaterally. However, in R Ml, younger adults show a moderate negative
BOLD response while the older adults show a large positive BOLD response.
Proceeding clockwise, L SMA and R SMA each showed moderate positive BOLD HRFs
for both age groups. For R Ml, however, younger adults had a moderately negative HRF
while older adults had a large positive HRF. Figure 4 confirms in group averages the
observations illustrated for individual subjects by Figures 2 and 3. Figure 5 compares
averaged Talairached younger and older groups' brains, showing HRF magnitude color-
coded across brain locations in a "glass brain" view
Results from a contiguity group repeated measures, multivariate ANOVA indicated
an interaction between age groups and area across the hemodynamic response function
(Hotelling's F(3, 10) = 8.345, p = .003). As this finding was consistent with our
hypothesis of differences in HRF across age groups and regions, we did not interpret
Figure 5: Group means overlayed on axial images for A) older adults and B) younger
adults during learned response execution after area-under-the-curve analysis
of estimated HRF. (Yellow indicates large positive area, orange indicates
moderate positive area, and blue indicates negative (below baseline) area.
The primary goal of this investigation was to compare negative BOLD responses
(NBR) in older and younger adults on a task previously shown to exhibit NBR. The main
findings are as follows. Individual subject analysis of maximally active voxels in the
primary (Ml) and supplementary motor (SMA) cortices showed different results for older
and younger adults during execution of learned button press sequences. All individuals
in each group evidenced typical positive BOLD responses (PBR) in Ml contralateral to
the response hand as well as in bilateral SMAs during response executions. Younger
participants showed the expected NBR in ipsilateral Ml while, surprisingly, all but one
older adult showed positive BOLD in this region. This difference in BOLD contrast
directions (NBR vs PBR) cannot be explained by performance differences, because the
age groups were nearly identical on measures of accuracy and reaction time for execution
of the learned button press sequence. These results indicate that healthy aging is linked
to changes in interhemispheric communication (see Peinnemann et al. 2002).
The present results are consistent with recent fMRI findings by Newton et al.
(2005). They studied 6 young adult participants (mean age 28) who were instructed to
execute or withhold a 4 Hertz thumb button press after a prepare cue while being scanned
in a block paradigm at 3T. Response execution versus passive fixation point viewing
elicited ipsilateral (to the response hand) Ml NBR with concomitant PBR in the
contralateral Ml. Newton et al. (2005) suggested that negative BOLD might signify
preparation of the ipsilateral Ml for potential cooperation with contralateral Ml if motor
processing became complex enough to require bilateral recruitment. Reduced activity in
ipsilateral Ml could serve to suppress its recruitment by other connections into this area.
The present results are also generally consistent with the model of hemispheric
asymmetry reduction in older adults, or HAROLD (Cabeza, 2002). The HAROLD model
evolved to explain an accumulation of neuroimaging studies comparing BOLD
activations of older and younger adults performing the same tasks, which showed with
some consistency that older adults have less lateralized activation than that seen in
younger adults (Rueter-Lorenz et al. 2000; Garavan et al. 1999; Dixon et al. 2004;
Cabeza et al. 2004). The HAROLD model has been used to describe aging related
activation patterns for a number of specific tasks involving the prefrontal cortex including
verbal working memory, visuo-spatial working memory and pre-potent response
inhibition. Support for the model has also accrued from investigations of hippocampus
(Maguire & Frith, 2003) and motor cortex (Naccarato et al. 2004; James, 2005; Baliz et
al. 2005). One explanation for the HAROLD phenomenon is compensation for declining
functionality in the aging brain by recruitment of cortical homologues (Cabeza, 2002;
Cabeza et al. 2004; Baliz et al. 2005; see also Salthouse & Babcock, 1991).
The Newton et al. (2005) suggestion that NBR in young adults might signify
preparation of R Ml for potential cooperation with L Ml should motor processing
became complex enough, combined with normally declining functionality in the aging
brain and HAROLD, would lead one to expect the BOLD signal direction change
(younger RM1 NBR to older RM1 PBR) observed in the present study. Our findings thus
add further support to the notion of neural compensation by recruitment of cortical
homologues in normal aging.
There are many plausible candidates to serve as the neurophysiological basis for
age-related decline in brain functionality. Age-related vascular changes are common
(Fang, 1976). Structural magnetic resonance imaging has shown that normal aging is
associated with a decrease in the thickness of the neocortex (Salat et al. 2005). Diffusion
tensor imaging has shown that aging is associated with significant increases of water
diffusion, probably resulting from decreases in the diffusion-limiting myelinated fibers
connecting cortical and/or subcortical areas (Pfefferbaum et al. 2005; Salat et al. 2005;
see also Madden et al. 2004). Dopamine uptake in caudate and putamen declines with
age as indexed by levels of striatal dopamine transporters (Kish et al. 1992; van Dyck et
al. 1995; van Dyck et al. 2002; Erixon-Lindroth et al. 2005). Evidence exists that
interhemispheric signaling could make use of cortico-striato-thalamo-cortical loops
(Gerloff et al. 1998; Kuhn et al. 2004; but see also Chen et al. 2004). As an example, a
recent case study reported that a patient lacking the corpus callosum showed TMS-
evoked reduction in ipsilateral motor evoked potentials, thus implicating subcortical
pathways in interhemispheric communication (McNair, 2004; though see Reddy et al.
The present study has two key limitations. First, locations of the NBRs in younger
adults were rather variable (see Table 2), but such variability is common for NBRs
(Allison et al. 2000; Nirkko et al. 2001; Born et al. 2002; Shmuel et al. 2002; Hamzei et
al. 2002; Aizenstein et al. 2004; Stefanovic et al. 2004; Stefanovic et al.. 2005; Newton et
al. 2005). Second, age group comparisons with such a small sample size (n=6 per group)
have questionable generality to the older population. Both Cabeza (2002) and Reuter-
Lorenz & Lustig (2005) emphasize inter-individual variability of functional imaging
results across older adults. This was apparent in the present study because one older
adult (s05, age 68, a champion marathon runner) showed NBR, not PBR, in R Ml during
response execution, a finding which was similar to the NBRs of younger adults. It is quite
likely that variations in social activity, exercise patterns, mental engagement, and other
factors, all contribute to the effects of normal aging on brain activations (Peinnemann et
al. 2002; Dixon et al. 2004). Both of these limitations could be mitigated by expanding
the sample size.
Implications and Future Directions
The current findings may have important implications for the focused development
of motor rehabilitation treatments because the age of the individual has an important
influence on how their primary motor cortical homologues interact. Both practitioners
and researchers are beginning to better grasp the manner in which the brain reorganizes
after stroke or other damage. A primary consideration for the development of treatment
regimens is the ability to identify substrates in the remaining viable tissue to which
functions formerly resident in the damaged tissue can migrate. The use of fMRI to
monitor brain plasticity during recovery of function provides a new tool for the continued
development of customized rehabilitation programs. The present evidence implies that
brain rehabilitation programs may need to be different for people of different age ranges.
LIST OF REFERENCES
Aizenstein, H. J., Clark, K. A., Butters, M. A., Cochran J. L., Stenger, V. A., Meltzer, C.
C., Reynolds, C. F., Carter, C. S., 2004. The BOLD hemodynamic response in
healthy aging. Journal of Cognitive Neuroscience 16, 786-793.
Alexander G.E., Crutcher M.D., DeLong M.R., 1990. Basal ganglia-thalamocortical
circuits, parallel substrates for motor, oculomotor, "prefrontal" and "limbic"
functions. Progress in Brain Research 85, 119-46.
Allison J.D., Meador K.J., Loring D.W., Figueroa R.E., Wright J.C., 2000. Functional
MRI cerebral activation and deactivation during finger movement. Neurology 54,
Amedi A., Malach R., Pascual-Leone A., 2005. Negative BOLD differentiates visual
imagery and perception. Neuron 48, 859-72.
Baliz, Y., Armatas, C., Farrow, M., Hoy, K. E., Fitzgerald, P. B., Bradshaw, J. L., and
Georgiou-Karistianis, N., 2005. The influence of attention and age on the
occurrence of mirror movements. Journal of the International Neuropsychological
Beaulieu C., Colonnier M. J., 1985. Number and size of neurons and synapses in the
motor cortex of cats raised in different environmental complexities. Comparative
Neurology 289, 178-81.
Born A.P., Law I., Lund T.E., Rostrup E., Hanson L.G., Wildschiodtz G., Lou H.C.,
Paulson O.B., 2002. Cortical deactivation induced by visual stimulation in human
slow-wave sleep. Neurolmage 17, 1325-35.
Buckner, R.L., Snyder, A.Z., Sanders, A.L., Raichle, M.E., Morris, J.C., 2000. Functional
brain imaging of young, nondemented, and demented older adults. Journal of
Cognitive Neuroscience 12, 24 -34.
Cabeza R., 2002. Hemispheric asymmetry reduction in older adults: the HAROLD
model. Psychological Aging 17, 85-100.
Cabeza R., Daselaar S.M., Dolcos F., Prince S.E., Budde M., Nyberg L., 2004. Task-
independent and task-specific age effects on brain activity during working memory,
visual attention and episodic retrieval. Cerebral Cortex 14, 364-75.
Chen C.C., Tyler C.W., Liu C.L., Wang Y.H., 2005. Lateral modulation of BOLD
activation in unstimulated regions of the human visual cortex. Neuroimage 24, 802-
Chen R., Yung, D., Li, J.Y., 2003. Organization of ipsilateral excitatory and inhibitory
pathways in the human motor cortex. Journal of Neurophysiology 89, 1256-64.
Dixon V., Kaufman, J.N., Antuono, P., Garavan, H., 2004. Experimental Evidence that
Age-Related Brain Activation Increases reflect a Recruitment Process. Poster
presented at Human Brain Mapping 2004, Toronto, Ontario.
D'Esposito M., Deouell L.Y., Gazzaley A., 2003. Alterations in the BOLD fMRI signal
with ageing and disease: a challenge for neuroimaging. Nature Reviews:
Neuroscience 4 863-72.
D'Esposito, M., Zarahn, E., Aguirre, G. K., and Rypma, B., 1999. The effect of normal
aging on the coupling of neural activity to the BOLD hemodynamic response.
Neurolmage 10, 6-14.
Erixon-Lindroth N, Farde L, Wahlin TB, Sovago J, Halldin C, Backman L., 2005. The
role of the striatal dopamine transporter in cognitive aging. Psychiatry Research
Fang, H.C.H., 1976. Observations on aging characteristics of cerebral blood vessels,
macroscopic and microscopic features, In S. Gerson and R.D. Terry, Eds..,
Neurobiology of Aging, Vol. 3, Raven, New York, 1976, 155-166.
Garavan, H., Ross, T. J., Stein, E. A., 1999. Right hemispheric dominance of inhibitory
control, An event-related functional MRI study. Proceedings of the National
Academy of Sciences, USA 96, 8301-8306
Gardner, J. L., Sun, P., Waggoner, R. A., Ueno K., Tanaka, K., and Cheng K., 2005.
Differences in temporal dynamics of positive and negative BOLD responses. Paper
presented at ISMRM 2005, Miami, FL.
Gerloff C., Richard J., Hadley J., Schulman AE, Honda M, Hallett M., 1998. Functional
coupling and regional activation of human cortical motor areas during simple,
internally paced and externally paced finger movements. Brain 121, 1513-31.
Hallett, M., 2000. Trancranial magnetic stimulation and the human brain. Nature 406,
Hamzei F., Dettmers C., Rzanny R., Liepert J., Buchel C., Weiller C., 2002. Reduction
of excitability ("inhibition") in the ipsilateral primary motor cortex is mirrored by
fMRI signal decreases. Neurolmage 17, 490-6.
Hamzei F., Liepert J., Dettmers C., Weiller C., Rijntjes M., In Press. Two different
reorganization patterns after rehabilitative therapy: An exploratory study with fMRI
and TMS. Neurolmage.
Hesselmann V., Zaro Weber O., Wedekind C., Krings T., Schulte O., Kugel H., Krug B.,
Klug N., Lackner K.J., 2001. Age related signal decrease in functional magnetic
resonance imaging during motor stimulation in humans. Neuroscience Letters 308,
Huettel S.A., Singerman J.D., McCarthy G., 2001. The effects of aging upon the
hemodynamic response measured by functional MRI. Neurolmage 13, 161-175.
Hummel F., Saur R, Lasogga S, Plewnia C, Erb M, Wildgruber D, Grodd W, Gerloff C.,
2004. To act or not to act. Neural correlates of executive control of learned motor
behavior. Neurolmage 23, 1391-401.
Hummel F., Andres F., Altenmuller E., Dichgans J. and Gerloff, C., 2002. Inhibitory
control of acquired motor programmes in the human brain. Brain 125, 404-420.
Jahanshahi M., Jenkins I.H., Brown R.G., Marsden C.D., Passingham R.E., Brooks D.J.,
1995. Self-initiated versus externally triggered movements. I. An investigation
using measurement of regional cerebral blood flow with PET and movement-
related potentials in normal and Parkinson's disease subjects. Brain 118, 913-33.
Kish S.J., Shannak K., Rajput A., Deck J.H., Homykiewicz 0., 1992. Aging produces a
specific pattern of striatal dopamine loss, implications for the etiology of idiopathic
Parkinson's disease. Journal of Neurochemistry 58, 642-8.
Ko6s T., Tepper, J.M., 1999. Inhibitory control of neostriatal projection neurons by
GABAergic intemeurons. Nature Neuroscience 2, 467-472.
Krampe, R. T., 2002. Aging, expertise and fine motor movement. Neuroscience:
Biobehavioral Review 26, 769-776.
Kuhn A.A., Brandt S.A., Kupsch A., Trottenberg T., Brocke J., Irlbacher K., Schneider
G.H., Meyer B.U., 2004. Comparison of motor effects following subcortical
electrical stimulation through electrodes in the globus pallidus internus and cortical
transcranial magnetic stimulation. Experimental Brain Research 155, 48-55.
Laurienti P.J., Field A.S., Burdette J.H., Maldjian J.A., Yen Y-F., Moody D.M., 2002.
Dietary caffeine consumption modulates fMRI measures. Neurolmage 17, 751-757
Laurienti P.J., Field A.S., Burdette J.H., Maldjian J.A., Yen Y.F., Moody D.M.,, 2003.
Relationship between caffeine-induced changes in resting cerebral perfusion and
blood oxygenation level-dependent signal. American Journal of Neuroradiology 24,
Liepert J., Dettmers C., Terborg C., Weiller C., 2001. Inhibition of ipsilateral motor
cortex during phasic generation of low force. Clinical Neurophysiology 112, 114-
Logothetis, N.K., Pauls, J., Augath, M., Trinath, T., Oeltermann, A., 2001.
Neurophysiological investigation of the basis of the fMRI signal. Nature 412, 150-
Madden D.J., Whiting W.L., Provenzale J.M., Huettel, S.A., 2004. Age-related changes
in neural activity during visual target detection measured by fMRI. Cerebral Cortex
Maguire, E.A., Frith, C.D., 2003. Aging affects the engagement of the hippocampus
during autobiographical memory retrieval. Brain 126, 1509-10.
McConnell K.A., Bohning D.E., Nahas Z., Shastri A., Teneback C., Lorberbaum J.P.,
Lomarev M.P., Vincent D.J., George M.S., 2003. BOLD fMRI response to direct
stimulation, transcranial magnetic stimulation of the motor cortex shows no decline
with age. Journal of Neural Transmission 110, 495-507.
McNair, N.A., 2004. Hemispheric interaction associated with motor-task complexity may
be mediated by subcortical pathways in acallosal individuals. Poster presented at
the Annual Meeting of the Cognitive Neuroscience Society 2004, San Francisco.
Naccarato, M., Calautti, C., Day, D.J., Fletcher, P.C., Jones, P.S., Carpenter, A.T.,
Bullmore, E.T., Baron, J., 2004. Does age affect hemispheric balance during finger
tapping? An fMRI study. Poster presented at Human Brain Mapping 2004,
Newton J.M., Sunderland A., Gowland P.A., 2005. fMRI signal decreases in ipsilateral
primary motor cortex during unilateral hand movements are related to duration and
side of movement. Neurolmage 24, 1080-7.
Nirkko, A.C., Ozdoba, C., Redmond, S.M.,Buerki, M., Schroth, G., Hess, C.W. and
Wiesendanger, W., 2001. Different Ipsilateral Representations for Distal and
Proximal Movements in the Sensorimotor Cortex, Activation and Deactivation
Patterns, Neurolmage 13, 825-835.
Ogawa S., Lee T.M., Kay A.R., Tank D.W., 1990. Brain magnetic resonance imaging
with contrast dependent on blood oxygenation. Proceedings of the National
Academy Sciences U.S.A. 87, 9868-72.
Peinemann A., Lehner C., Conrad B., Siebner H.R., 2001. Age-related decrease in paired-
pulse intracortical inhibition in the human primary motor cortex. Neuroscience
Letters 313, 33-6.
Pfefferbaum A., Adalsteinsson E., Sullivan E.V., 2005. Frontal circuitry degradation
marks healthy adult aging: Evidence from diffusion tensor imaging. Neurolmage
Rao S.M., Harrington, Haaland K., Bobholz J.A., Cox R.W. and Binder J.R., 1997.
Distributed neural systems underlying the timing of movements. Journal of
Neuroscience 15, 5528-35.
Reuter-Lorenz, P. A., Jonides, J., Smith, E. S., Hartley, A., Miller, A., Marshuetz, C.,
2000. Age differences in the frontal lateralization of verbal and spatial working
memory revealed by PET. Journal of Cognitive Neuroscience 12, 174-187.
Reddy H., Lassonde M., Bemasconi A., Bemasconi N., Matthews P.M., Andermann F.,
Amold D.L., 2000. An fMRI study of the lateralization of motor cortex activation
in acallosal patients. Neuroreport 11, 2409-13.
Reuter-Lorenz, P. A., and Lustig, C., 2005. Brain aging: reorganizing discoveries about
the aging mind. Current Opinion in Neurobiology 15, 245-251.
Ridding M.C., Brouwer B., and Nordstrom M.A., 2000. Reduced interhemispheric
inhibition in musicians. Experimental Brain Research 133, 249-253.
Roether J., Knab R., Hamzei F., Fiehler J., Reichenbach J.R., Buchel C., Weiller C.,
2002. Negative dip in BOLD fMRI is caused by blood flow--oxygen consumption
uncoupling in humans. Neurolmage 15, 98-102.
Ross M.H., Yurgelun-Todd D.A., Renshaw P.F., Maas L.C., Mendelson J.H., Mello
N.K., Cohen B.M., Levin J.M., 1997. Age-related reduction in functional MRI
response to photic stimulation. Neurology 48, 173-6.
Salat D.H., Tuch D.S., Greve D.N., van der Kouwe A.J., Hevelone N.D., Zaleta A.K.,
Rosen B.R., Fischl B., Corkin S., Rosas H.D., Dale A.M., 2005. Age-related
alterations in white matter microstructure measured by diffusion tensor imaging.
Neurobiol Aging 26, 1215-27.
Salthouse, T.A., Babcock, R.L., 1991. Decomposing adult age differences in working
memory. Developmental Psychology 27, 763-776.
Shmuel A., Yacoub E., Pfeuffer J., Van de Moortele P.F., Adriany G., Hu X., Ugurbil K.,
2002. Sustained negative BOLD, blood flow and oxygen consumption response
and its coupling to the positive response in the human brain. Neuron 19, 1195-210.
Stefanovic, B., Wamking, J. M., and Pike, G. B., 2004. Hemodynamic and metabolic
responses to neuronal inhibition. Neurolmage 22, 771-778.
Stefanovic, B., Wamking, J. M., Kobayashi, E., Bagshaw, A. P., Hawco, C., Dubeau, F.,
Gotman, J., and Pike, G. B., 2005. Hemodynamic and metabolic responses to
activation, deactivation and epileptic discharges. Neurolmage 28, 205-215.
van Dyck C.H., Seibyl J.P., Malison R.T., Laruelle M., Wallace E., Zoghbi S.S., Zea-
Ponce Y., Baldwin RM., Charney D.S., Hoffer P.B., 1995. Age-related decline in
striatal dopamine transporter binding with iodine-123-beta-CITSPECT. Journal of
Nuclear Medicine 36, 1175-81.
van Dyck C.H., Seibyl J.P., Malison R.T., Laruelle M., Zoghbi S.S., Baldwin R.M., Innis
R.B., 2002. Age-related decline in dopamine transporters, analysis of striatal
subregions, nonlinear effects, and hemispheric asymmetries. American Journal of
Geriatric Psychiatry 10, 36-43.
Verstynen T., Diedrichsen J., Albert N., Aparicio P., Ivry R.B., 2005. Ipsilateral motor
cortex activity during unimanual hand movements relates to task complexity.
Journal of Neurophysiology 93, 1209-22.
Ward, B.D., 2002. 3dDeconvolve. Retrieved March 21, 2006, from National Institute of
Health Web site: http://afni.nimh.nih.gov/pub/dist/doc/manual/Deconvolvem.pdf
Keith Matthew McGregor was born in Norwood, MA. After his 1997 graduation
from the College of the Holy Cross in Worcester, MA, with a B.A. in psychology, he
began a career in information technology. After working in the technology industry for a
number of years, Mr. McGregor returned to graduate school in 2003 at the University of
Florida entering a PhD program in cognitive psychology. He is married and currently
resides in Gainesville, FL.