Brain Oral Session: Functional Brain Imaging

602. Spatiotemporal complexity in the haemodynamic response to somatosensory stimulation in the un-anaesthetized rat

C. Martin¹, J. Berwick², A. Kennerley², Y. Zheng² and J. Mayhew²

¹Department of Physiology, Anatomy and Genetics, University of Oxford, Oxford; ²Department of Psychology, The University of Sheffield, Sheffield, UK

Background and aims: Blood oxygen level dependent (BOLD) signals are used in functional magnetic resonance imaging (fMRI) to localize and quantify task-dependent changes in brain activity. As the BOLD signal is an indirect marker of neural activity, characterization of both the spatial and temporal haemodynamic response functions (changes in cerebral blood flow, volume and oxygen consumption) are critical for the accurate interpretation of functional imaging data. Our goal in this study was to use optical imaging spectroscopy (OIS) in a previously developed un-anesthetized rat model to extend current understanding of the spatiotemporal haemodynamic changes that underlie both positive and negative BOLD signals. OIS measures changes in the spectrum of remitted light from brain tissue which can be compared to known absorption spectra in order to calculate changes in haemoglobin concentration and oxygenation.

Methods: Under surgical anaesthesia the skull was exposed and a section overlying somatosensory cortex was thinned to translucency. An imaging chamber was then placed over the thinned region of skull and secured with skull screws and dental cement. To deliver electrical stimulation to the awake animals, Teflon coated tungsten microwires were chronically implanted into the contralateral whisker pad and fed subcutaneously to a connector adjacent to the imaging chamber. The animals were treated with an analgesic and left to recover for 3 to 5 days. For each imaging experiment, trained animals were placed into a harness and to reduce head movements, a pneumatically operated clamp secured the implanted imaging chamber and therefore the head. A medical endoscope was used to provide both illumination of the cortex and transmission of the remitted images to a digital camera. Stimulation consisted of a 16-second, 5 Hz pulse train (0.4 mA) with an individual pulse width of 0.3 ms. The spectral analysis of remitted images produced 2-D maps, over time, of changes in oxyhaemoglobin (HbO₂), deoxyhaemoglobin (Hbr) and total haemoglobin (HbT) concentration (A).

Results and conclusions: The main finding of this study was that monotonic stimulation of the contralateral whisker pad in awake rats led to a complex spatiotemporal haemodynamic response in cortex. In addition to regions exhibiting a haemodynamic response predictive of a positive BOLD signal (increased HbO₂ and HbT, decreased Hbr—A and B), large regions of somatosensory cortex displayed an inverted haemodynamic response (decreased HbO₂ and HbT, increased Hbr) that is predictive of a negative BOLD signal (A and C). The magnitude and spatial extent of these signals varied during the stimulation period and furthermore, some regions of cortex were characterised by ‘biphasic’ haemodynamic changes, where initial signal increases in response to stimulation onset were followed by below baseline signal decreases as stimulation continued (C). These findings may have important implications for our understanding of the haemodynamic changes that underlie positive and negative BOLD signals.

301. Interhemispheric functional connectivity changes after transient focal ischemia in rat brain: a serial resting-state fMRI study

M.P.A. Van Meer^1,2, K. Van Der Marel¹, K. Wang³, W.M. Otte^1,2, J.W. Berkelbach Van Der Sprenkel² and R.M. Dijkhuizen¹

¹Image Sciences Institute, University Medical Center Utrecht; ²Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, Utrecht, The Netherlands; ³Institute of Automation, Chinese Academy of Sciences, Beijing, China

Objectives: Resting-state fMRI can identify the synchronization of spontaneous low-frequency fluctuations of the blood oxygenation level-dependent (BOLD) signal in the brain, indicative of functional connectivity (FC). ¹ Because no peripheral stimulation is required, resting-state fMRI is an ideal tool to study dynamics of changes in functional organization of different neuronal networks after brain injury. In this study we applied serial resting-state fMRI to assess changes within the sensorimotor network after unilateral stroke in rats.

Methods: Experimental stroke was induced by 90-min transient intraluminal occlusion of the right middle cerebral artery (tMCA-O) in adult male rats (n = 14). Sensorimotor function was measured longitudinally by measuring neurological deficiency score (NDS) and adhesive removal time from the affected forelimb. ² Structural MRI (T₂-weighted) and resting-state BOLD fMRI (T₂*-weighted gradient echo EPI) measurements were acquired before and at 3, 7, 21, and 70 days after stroke on a 4.7T MR system. Rats were ventilated with 1 to 2% isoflurane in air/O₂ (2:1). Animals were divided in two groups based lesion extent: Group I animals had only subcortical lesions (n = 5); Group II animals had both cortical and subcortical lesions (n = 9). Low-frequency BOLD fluctuations were band-pass filtered between 0.01 and 0.08 Hz. Lesioned tissue was excluded from the analysis. Interhemispheric FC was measured as the correlation coefficient r between low-frequency BOLD signals in ipsi- and contralateral S1fl (forelimb region of the primary somatosensory cortex) and its Fisher-transformed z′-value.

Results: The figure shows the temporal pattern of mean (±SD) interhemispheric FC between ipsi- and contralateral S1fl. At 3 days after stroke, interhemispheric FC was reduced in both groups. Group I animals demonstrated recovery of interhemispheric FC at 7 days after tMCA-O, which was accompanied by an improvement of neurological function and adhesive removal time. In Group II animals, however, interhemispheric FC remained reduced at 7 days, in parallel with lower sensorimotor function scores as compared to Group I animals. Partial recovery of interhemispheric FC in Group II was observed after 10 weeks, when behavioral scores had also improved.

Figure

Rest-state.

Conclusions: This study shows that resting-state fMRI can be applied to study changes in FC in a rat stroke model. Acutely after stroke a decrease was detected in interhemispheric FC of bilateral S1fl, despite the absence of structural ischemic damage in the ipsilesional S1fl. Animals with only subcortical ischemic lesions showed relatively fast restoration of interhemispheric FC, whereas animals with both cortical and subcortical lesions showed delayed recovery. We found correspondence between the evolution of FC within the bilateral sensorimotor network and changes in sensorimotor scores, which suggests that resting-state fMRI provides a valuable method to study brain reorganisation in relation to functional recovery after cerebral injury.

799. Bold impulse responses for event related and steady-state paradigms: implications for CMR_O2

P. Herman^1,2,3, B.G. Sanganahalli^1,2, H. Blumenfeld^2,4,5,6 and F. Hyder^1,2,7

¹Department of Diagnostic Radiology; ²Quantitative Neuroscience with Magnetic Resonance, Yale University, New Haven, Connecticut, USA; ³Institute of Human Physiology and Clinical Experimental Research, Semmelweis University, Budapest, Hungary; ⁴Department of Neurology; ⁵Department of Neurosurgery; ⁶Department of Neurobiology; ⁷Department of Biomedical Engineering, Yale University, New Haven, Connecticut, USA

Objectives: Quantitative mapping of changes in cerebral oxygen consumption with calibrated fMRI has become a popular modality for studying functional brain activity ¹ because it is proportional to changes in energy consumption associated with alterations in neuronal activity induced by the stimulation. ² This approach is based on a model which describes tissue oxygen extraction at steady-state,^3–4 and this model is not proved to use for dynamic calculation. Experimental demonstration of linearity between neural and BOLD-related responses to short and long-lasting stimulations is needed to achieve an empirical proof of the feasibility of the steady-state BOLD model to dynamic events, then it may be possible to use calibrated fMRI in a dynamic manner.

Methods: Sprague-Dawley rats were anesthetized with Halothane or Isoflurane which switched to i.p. α-chloralose after the surgery. Forepaw Stimulation: Each stimulus train used 2 mA in amplitude and 0.3 ms in duration with 3 Hz frequency. The stimulus number of the event related stimuli was varied from 1 to 4. 90 pulses train was used for steady-state stimulation. BOLD (n = 12): All gradient echo EPI fMRI data were obtained on a modified 11.74T Bruker horizontal-bore spectrometer (Billerica, MA) using a ¹H resonator/surface coil RF probe. Electrophysiology (n = 14): A separate group of animals were subject of the electrophysiological measurement with microelectrodes above the forepaw somatosensory region. Local field potentials (LFP) were obtained applying low pass filter (<150 Hz) to the raw time series. Convolution analysis: The transfer function between the LFP and the BOLD signal was modeled by the gamma variate function. ⁵ The parameters of the transfer function were calculated with iterative steps. The input function was defined as the average of the LFP series. The input signal covered not only one stimulus response, but consecutively all of the short stimulus conditions. The individual events were normalized to the largest evoked potential.

Results: The same transfer function was used not only for modeling event-related responses, but for long lasting stimulation events too. The precision of the simulated signal was checked by the comparison of the residual signal (i.e., difference between the modeled and measured signal) and the standard deviation of the measured signal. We considered the precision of the model adequate since the root mean square of the residual signal was smaller then the mean standard deviation of the measured signal. We got similar results from other modalities (CBV and CBF) (data not shown).

Conclusion: We showed that the transfer functions of fMRI responses generated by convolution analysis with neural activity are time invariant for events in the second to minute range, therefore the steady-state BOLD model ³ may be used for dynamic CMR_O2 calculation.

This work was supported by grants from NIH (R01 MH-067528, R01 DC-003710, P30 NS-52519).

490. Functional bolus-tracking arterial spin labeling: a new approach to quantitative fMRI

M. Kelly¹, C. Blau¹, O. Gobbo¹, K. Griffin², J. Jones² and C. Kerskens¹

¹Trinity College Institute of Neuroscience, Trinity College Dublin; ²School of Medicine, University College Dublin, Dublin, Ireland

Objectives: The alterations in the BOLD signal when cerebral blood flow, volume and vascular structure change (due to aging and or neurovascular disease) are relatively unknown. This presents a considerable obstacle when interpreting BOLD fMRI studies. ¹ The purpose of this study is to develop a new quantitative fMRI technique, bolus-tracking arterial spin labelling (ASL) fMRI, which provides a quantitative assessment of the blood perfusion of an activated brain region. The technique was used to quantify changes in both the mean and capillary transit time (MTT and CTT) during neuronal activation in the rat brain.

Methods: A Fokker-Planck equation that describes the distribution of labelled arterial water in the brain during ASL experiments was derived. ² A bolus-tracking ASL sequence was designed to provide concentration-time curves that represent the passage of a bolus of labelled arterial water through the ROI (graph). Wistar rats (n = 6) were sedated with medetomidine, which has been shown to provide suitable conditions for fMRI studies in rats (Weber et al, 2006). Electrical stimulation of the right forepaw resulted in neuronal activation in the left primary somatosensory cortex forelimb (S1FL) region.

The left and right S1FL regions were selected as the activated and control ROI respectively. The solution to the Fokker-Planck equation for the boundary conditions describing the bolus-tracking ASL experiment was fitted to the concentration-time curves (graph). The MTT and CTT were calculated from the first and second moments of the resultant curve respectively.^3,4

Results: The graph shows the concentration-time curves fitted to the Fokker-Planck model for the control and activated ROIs. The mean MTT and CTT were 1.87s±0.18s and 1.65s±0.06s respectively for the control ROI. For the activation ROI, the mean MTT and CTT were reduced to 1.68s+0.2s and 1.37s±0.17s respectively. This represents a statistically significant difference (P<0.01) bet;ween the activated and control ROI for both transit times.

Figure

Modelfitted to data for activated and control ROI.

Conclusions: We have developed a new fMRI technique that quantifies the change in MTT and CTT during neuronal activation. The measured decrease in MTT during neuronal activation was expected, as flow to an activated region is known to increase, MTT is proportional to 1/Flow and CTT is proportional to 1/Flow. ² The technique offers the potential to longitudinally monitor changes in these parameters during neuronal activation due to increasing age or disease progression.

261. HIV causes premature aging of brain function

B. Ances¹, F. Vaida², J. Yeh³, I. Grant⁴, A. Mccutchan³, R. Ellis⁵ and R. Buxton⁶

¹Neurology, Washington University School of Medicine, Saint Louis, Missouri; ²Family and Preventative Medicine; ³Medicine; ⁴Psychiatry; ⁵Neurosciences; ⁶Radiology, University of California at San Diego, San Diego, California, USA

Background and aims: Highly active antiretroviral therapy (HAART) has increased the prevalence of HIV infected older individuals (>50 years old). A growing need has arisen to characterize the effects of aging and HIV infection. While the effects of aging and HIV have been studied in other organ systems, their interaction on brain function has not been investigated. We used non-invasive functional magnetic resonance imaging (fMRI) to measure the blood oxygen level dependent (BOLD) effect, cerebral blood flow (CBF), and calculated cerebral metabolic rate of oxygen consumption (CMRO₂) within HIV infected (HIV+) and HIV− controls. We hypothesized that HIV infection will confer a proaging effect on brain function due to ongoing viral induced inflammation and oxidative stress.

Methods: BOLD and CBF responses were obtained from 26 HIV+ subjects and 25 HIV− controls, from 20 to 62 years old, on a 3 Tesla General Electric scanner. All subjects underwent both mild hypercapnia and functional activation experiments. Mild hypercapnia provided a calibration method for calculating functional CMRO₂ changes. Functional activation consisted of a black and white radial checkerboard flickering at 8 Hz. Clusters of CBF activated voxels within the visual cortex (VC) were assessed. A Wilcoxon rank test investigated if differences in measured fMRI outcomes (baseline CBF, functional changes in CBF, BOLD, and CMRO₂) were present between HIV+ and HIV− subjects. The association and the interaction between fMRI measures, HIV status, and age were investigated using a multiple regression model.

Results: The median age for HIV+ subjects and HIV− controls was similar with no significant differences observed for either sex or education. Baseline CBF was reduced for both age (P<0.001) and HIV status (P<0.001). HIV infection was equivalent to a 15 year increase in age. Similarly, both age (P = 0.005) and HIV (P = 0.01) led to increases in functional CBF changes, with HIV infection equivalent to a 21 year increase in age. Functional BOLD changes decreased with age (P = 0.001) and HIV status (P = 0.05). HIV infection was equivalent to a 15 year increase in age dor functional BOLD changes. Calculated functional CMRO₂ changes (P<0.001) increased with both age (P<0.001) and HIV status (P = 0.007) with a significant interaction present (P = 0.001).

Conclusions: Our results suggest that HIV infection accelerates aging effects on brain function. We hypothesize that these changes in brain function could be caused by ongoing inflammation and oxidative stress due to HIV. Functional CBF and CMRO₂ changes were elevated in HIV+ subjects compared to HIV− controls, possibly suggesting an element of increased metabolic requirements due to neuroinflammation. However, the coupling between CBF and CMRO₂ remained preserved suggesting that HIV infection does not alter the basic physiology between increased metabolic requirements and oxygen delivery.

245. Contribution of dopamine D1 and D2 receptors to amygdala activity

H. Takahashi, T. Otsuka, H. Takano, F. Kodaka, H. Kikyo, R. Arakawa, M. Miyoshi, M. Okumura, H. Ito and T. Suhara

National Institute of Radiological Sciences, Chiba, Japan

Objectives: To investigate the contribution of dopamine receptor subtypes to amygdala activation in human, we conducted a multimodal in vivo neuroimaging study in which dopamine D1 and D2 receptor bindings in the amygdala were measured with positron emission tomography (PET) and amygdala activation in response to fearful faces was assessed by functional magnetic resonance imaging (fMRI) in healthy volunteers.

Methods: Fourteen healthy male volunteers were studied. PET studies were performed on ECAT EXACT HR+. For evaluation of dopamine D1 and D2 receptors, [¹¹C]SCH23390 and [¹¹C]FLB457 were used, respectively. Dynamic scans were performed for 60.0 min for [¹¹C]SCH23390 and 90.0 min for [¹¹C]FLB457. Quantitative analysis was performed using the three-parameter simplified reference tissue model. The cerebellum was used as a reference region. Parametric images of BP_ND of [¹¹C]SCH23390 and [¹¹C]FLB457 were created. Amygdala activations in response to fearful faces were examined using fMRI. The fMRI were acquired with a 3.0 Tesla system. Data analysis was performed with the statistical parametric mapping software (SPM2). To conduct multimodality voxel-wise correlation analysis between BOLD signal and dopamine receptor binding, we used the biological parametric mapping (BPM) toolbox for SPM.

Results: Voxel-wise correlation analysis by BPM revealed that dopamine D1 binding in the amygdala was positively correlated with amygdala activation in response to fearful faces (R>0.6, P<0.015), but dopamine D2 binding in the amygdala was not related to amygdala activation.

Conclusion: Dopamine D1 receptors might play a more vital role in enhancing amygdala response than dopamine D2 receptors when sensory inputs are affective.