Cerebral Cortex Advance Access originally published online on October 19, 2007
Cerebral Cortex 2008 18(6):1314-1325; doi:10.1093/cercor/bhm163
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Sensory Inputs from Whisking Movements Modify Cortical Whisker Maps Visualized with Functional Magnetic Resonance Imaging
1 MRC Centre for Neurodegeneration Research, King's College London, DeCrespigny Park, London SE5 8AF, UK, 2 Department of Mechanical Engineering, Imperial College London, London SW7 2AZ, UK, 3 Division of Engineering, King's College London, Strand, London WC2R 2LS, UK, 4 Current address: Institute of Cancer Studies, Department of Medicine, University College London, London WC1E 6JJ, UK
Address correspondence to email: g.finnerty{at}iop.kcl.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionFundingReferences |
|---|
Rodents vary the frequency of whisking movements during exploratory and discriminatory behaviors. The effect of whisking frequency on whisker cortical maps was investigated by simulating whisking at physiological frequencies and imaging the whisker representations with blood oxygen level–dependent (BOLD) functional magnetic resonance imaging. Repetitive deflection of many right-sided whiskers at 10 Hz evoked a positive BOLD response that extended across contralateral primary somatosensory cortex (SI) and secondary somatosensory cortex (SII). In contrast, synchronous deflection of 2 adjacent whiskers (right C1 and C2) at 10 Hz evoked separate positive BOLD responses in contralateral SI and SII that were predominantly located in upper cortical layers. The positive BOLD responses were separated and partially surrounded by a negative BOLD response that was mainly in lower cortical layers. Two-whisker representations varied with the frequency of simulated whisking. Positive BOLD responses were largest with 7-Hz deflection. Negative BOLD responses were robust at 10 Hz but were weaker or absent with 7-Hz or 3-Hz deflection. Our findings suggest that sensory inputs attributable to the frequency of whisking movements modify whisker cortical representations.
Key Words: cortex • fMRI • negative BOLD • rat • somatosensory • whiskers
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingReferences |
|---|
Much of the cortex is given over to maps of sensory inputs. These cortical maps are not static representations of the sensory periphery but are dynamic. For example, stimulation of one whisker evokes firing that is first recorded in the barrel column of the stimulated whisker. Neural activity then propagates across multiple barrel columns over tens of milliseconds (Moore and Nelson 1998
; Zhu and Connors 1999; Brecht et al. 2003) and can involve the whole barrel field (Petersen et al. 2003). It has been suggested that this cross-columnar signaling may be important for sensory processing during active touch (Carvell and Simons 1995).
Rats move their whiskers back and forth (commonly termed whisking) at frequencies of 4–12 Hz when investigating their environment (Welker 1964; Carvell and Simons 1990). The frequency of whisking is not random but can be modulated during active touch. Typically, exploratory behaviors are associated with whisking at 6–9 Hz, whereas discriminatory behaviors involve whisking at frequencies above 9 Hz (Carvell and Simons 1995; Harvey et al. 2001). It has been proposed that whisker representations change with the sensory task that is being performed. The hypothesis is that the whisker sensory system is set to detect contact when whiskers are stationary but is retuned by whisking behavior (Fanselow and Nicolelis 1999; Moore 2004). In primary somatosensory cortex (SI), whisker representations become more focused with alertness (Castro-Alamancos 2004) and with whisker deflection in the upper part of the whisking frequency range (Kleinfeld and Delaney 1996; Sheth et al. 1998). Furthermore, the frequency of whisking modifies the response in SI evoked by passive whisker stimulation (Fanselow and Nicolelis 1999; Crochet and Petersen 2006; Ferezou et al. 2006). These findings suggest that whisker representations exhibit frequency-dependent modifications.
Several techniques have been used to image whisker representations evoked by repetitive deflection. Intrinsic signal imaging (Masino et al. 1993; Devor et al. 2005) and voltage sensitive dye-based imaging (Kleinfeld and Delaney 1996; Petersen et al. 2003) give high-resolution images of whisker representations in SI. However, both techniques favor signals from upper cortical layers. Electrophysiological recordings indicate that individual cortical layers may behave differently because artificially induced whisking at 5 Hz evokes neuronal depression in cortical layer 2/3 and neuronal activation in layer 5a (Derdikman et al. 2006).
Functional magnetic resonance imaging (fMRI) with contrast based on blood oxygen level–dependent (BOLD) signals (Ogawa et al. 1990), cerebral blood flow (Kwong et al. 1992; Williams et al. 1992), or cerebral blood volume (CBV) (Mandeville et al. 1998) offers several approaches for depth-independent imaging of whisker maps. Here, we use BOLD fMRI to image whisker representations. We simulated whisking at physiological frequencies in anesthetized rats to isolate the sensory input attributable to large-amplitude whisker movements from the effects of inputs from whisker-related motor centers (Fee et al. 1997; Hentschke et al. 2006). Neural activity in upper layers of SI adapts with a few cycles of whisking (Moore 2004). Therefore, we aimed to image "steady-state" representations evoked by simulated whisking at different frequencies. We show that whisker representations imaged with BOLD fMRI vary with the number of deflected whiskers and exhibit frequency-dependent changes in the physiological whisking range. The results suggest that the frequency of whisking movements modifies whisker cortical maps.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingReferences |
|---|
Animal Preparation
All procedures were carried out in accordance with the UK Animals (Scientific Procedures) Act 1986. Anesthesia of male Sprague–Dawley rats (200–250 g) was induced with 4% isoflurane in 50% O2 and maintained initially with 2% isoflurane in a 1:1 mixture of O2:air for 5–10 min. A tail vein was cannulated, a bolus of 
-chloralose (65 mg/kg) was injected, and an infusion of -chloralose (30 mg/kg/h) started. Isoflurane anesthesia was discontinued 3 min after the bolus injection. The rat's head was restrained in a custom-made stereotaxic frame, which was isolated from the whisker stimulators to minimize head motion artifacts. Respiratory rate (prescanning, 73 ± 1 breaths per min; postscanning [3 h later] 65 ± 1 breaths per min, n = 6 animals) and rectal temperature were monitored during scanning (SA Instruments Inc., Stony Brook, NY). Rectal temperature was maintained at 37.0 ± 0.5 °C by a temperature-controlled hot air supply (SA Instruments Inc.). We restricted image acquisition to a 2.25 h period commencing 45 min after the -chloralose bolus was given to ensure that anesthesia was stable during scanning (Austin et al. 2005).
Whisker Stimulation
Rats have 5 rows of whiskers commonly denoted A–E (Welker and Woolsey 1974). The 2-whisker protocol comprised trimming all whiskers except for the 2 most caudal whiskers in the C row (C1 and C2) acutely to the level of the facial hair on both sides of the rat's snout (Fig. 1A). The right C1 and C2 whiskers were placed in the teeth of a comb located 1 cm away from the rat's snout (Fig. 1B). The multiple whisker–row protocol comprised placing 3–4 whiskers from each of the B–E rows into the comb. The remaining rostral whiskers in these rows were too short to reach the comb. The A row of whiskers and the outlying , β, 
, and 
whiskers were cut to prevent the comb from intermittently contacting them (Fig. 2B).
|
|
The comb was moved rostrocaudally at 3, 7, or 10 Hz (see Magnetic Resonance Imaging and fMRI methods for details of randomized block design used for experiments) by a pneumatic system controlled by a custom-written program (Labview, National Instruments, Austin, TX). We refer to the whisker displacement device as an actuator. The operation of the actuator was videoed outside the scanner with a high-speed camera (Phantom IV, Photo-sonics Inc., CA) at 1000 frames/s (pixel, 0.16 x 0.16 mm in specimen plane) to measure the kinetics of whisker deflection. During a deflection cycle, whisker displacement was 2.7 mm rostrocaudally with a maximum velocity of 0.15 m/s. The kinetics of the actuator during a deflection cycle were similar when operated at 3, 7, and 10 Hz. Operation of the actuator did not induce susceptibility artifacts when a phantom was imaged with our usual setup. Analysis of the phantom data set did not introduce artifactual negative or positive BOLD responses into our results.
Magnetic Resonance Imaging and fMRI Methods
Imaging was performed in a horizontal bore 9.4T magnet (Oxford Instruments, Oxford, UK) interfaced with a Varian Inova console (Varian, Palo Alto, CA) and equipped with a gradient set (ID 130 mm, 210 mT/m, 160-µs rise time). A 25-mm diameter surface coil (Varian) was used to transmit and receive the radiofrequency signal. The junction of the optic nerves at the optic chiasm, which is approximately at bregma along the rostrocaudal axis (Paxinos and Watson 1982), was used as an anatomical landmark in our scout scans. This served to determine the rostrocaudal location of the C1 and C2 whisker barrels columns, which are 2–3.5 mm caudal to bregma (Chapin and Lin 1984; Remple et al. 2003; Benison et al. 2007). Coronal anatomical scans were created from a spin-echo sequence (time repetition [TR]/time echo [TE] = 1000/20 ms) with 0.5-mm thick slices, a field of view (FOV) of 32 x 32 mm, a matrix size of 192 x 192, and 4 signals averaged. The 0.5-mm slice thickness approximates to the diameter of one-barrel column in SI.
Experiments that compared the 2-whisker protocol with the multiple whisker–row protocol in the same animal comprised one scan period divided into 2 contiguous sessions. One session compared 10-Hz deflection of the right C1 and C2 whiskers with no whisker deflection. The other session compared 10-Hz deflection of whiskers in the B–E rows on the left side of the snout with no whisker deflection. The order of the protocols during experiments was randomized. Experiments investigating the effect of deflection frequency were also divided into 2 contiguous sessions. The first session compared no whisker movement with deflection of the right C1 and C2 whiskers at one frequency chosen from 3, 7, and 10 Hz. The second session used one of the 2 remaining deflection frequencies for ON blocks. This protocol was favored over randomization of all 3 stimulation frequencies for each rat to ensure adequate averaging of images within the time constraint for stable anesthesia (Austin et al. 2005). These animals were used to prepare the group maps in Figure 6. An additional 4 data sets comprising 3-Hz and 7-Hz whisker deflection in randomized order were acquired to reduce the difference in number of animals tested with 3 Hz or 7 Hz compared with 10 Hz, thereby improving the power of statistical tests performed on single-animal data. Mock fMRI sequences were run for both 2-whisker and multiple whisker–row protocols prior to putting the rat into the scanner to ensure that the actuator did not touch whiskers too short to reach the comb or facial hairs during movement and that deflection of whiskers in the actuator comb did not cause the whiskers to pull on their respective follicles and cause pain.
The fMRI experiments used a randomized block design. Both sessions in an experiment comprised 120 blocks with 60 blocks ON (whisker movement continues throughout block) and 60 blocks OFF (no whisker movement throughout block). A volume of data was acquired throughout each block. Each volume comprised 12 slices of 0.5 mm thickness. A multiecho gradient echo imaging sequence was custom written for fMRI experiments to improve the contrast to noise of the BOLD signal (Posse et al. 1999). Imaging parameters were flip angle, 31 degrees; TR = 340 ms; TE = 4, 8, 12, 16, 20 ms; FOV, 32 x 32 mm; matrix, 96 x 96; acquisition time per volume, 32.6 s; and total scan time per fMRI session, 1 h 5 min. ON and OFF periods were coordinated with the scanning protocol by a TTL pulse sent from the imaging console to the Labview program. Pulsations due to the cardiac or respiratory cycles introduce colored noise into functional images, which was reduced by randomization of the ON and OFF blocks with the conditions that 1) the protocol always begins with an OFF and 2) a maximum of 2 ONs occur in succession.
Data Analysis
All fMRI data were converted from the Varian to ANALYZE format using custom-written software. Multiecho gradient echo data sets were converted into single-echo images (effective TE = 12 ms) by summing the magnitude images generated from each echo (Posse et al. 1999). Images from each scan session, which form a time series, were coregistered and aligned using the rigid body transformation algorithm in SPM99 (http://www.fil.ion.ucl.ac.uk/spm/). This corrects for motion along 3 translation and 3 rotation axes. A precise mask was applied to delineate the brain in the realigned data. Any rat that exhibited head motion exceeding 0.5 mm along x or y axes or more than 0.75 mm in the z direction (along B0) was excluded from further analysis.
We minimized BOLD signal attributable to large draining veins and vascular inflow (Menon and Goodyear 2001) by constructing a coefficient of variation map of the BOLD signal and eliminating voxels with coefficients of variation greater than 15% (Hlustik et al. 1998).
The BOLD signal recorded from the brain commonly includes structured (colored) noise, for example, due to pulsations attributable to the cardiac or respiratory cycles, and nonstructured (white) noise, which reduce BOLD contrast. We reduced noise in our functional images by performing a probabilistic independent component analysis on 4D data sets using MELODIC 2.0 (http://www.fmrib.ox.ac.uk/fsl/) (Thomas et al. 2002). The resulting components were compared with the stimulus paradigm using a Pearson correlation test. Components that had a correlation coefficient with a P value >0.1, that is, components were not correlated with the stimulus paradigm, were removed by linear regression to form a denoised data set. Further data processing and analysis were performed in SPM. Data were smoothed with a Gaussian kernel to conform to the expectation in SPM that the data approximate a random Gaussian field (Worsley and Friston 1995). The full-width half-maximum of the kernel (0.99 mm) was chosen to approximate the distributions of the electrophysiological and hemodynamic responses evoked by whisker deflection in SI and SII, thereby, optimize signal detection (Worsley et al. 1996; Triantafyllou et al. 2006) in single-animal data sets.
The BOLD signal in scalp muscle overlying somatosensory cortex fulfilled 2 roles in our experiments. First, it was examined to ensure that movement of the actuator had not induced artifactual BOLD signals, which were not found. Second, we used the mean BOLD signal within scalp muscle as an estimate of global effects during experiments (Shoaib et al. 2004; Lowe et al. 2007). The mean scalp muscle BOLD signal and the stimulation paradigm were independent at all frequencies (mean probability for null hypothesis that correlation coefficient = 0; 3 Hz, 0.47 ± 0.08; 7 Hz, 0.50 ± 0.10; 10 Hz, 0.41 ± 0.09; n = 8 for each group, P = 0.790, 1-way analysis of variance [ANOVA]). Therefore, we included both the mean BOLD signal in scalp muscle and the motion parameters from the head motion correction (Friston et al. 1996) as covariates of no interest in the statistical model of the BOLD signal. Addition of scalp muscle BOLD signal to statistical models did not introduce artifactual separation of the positive BOLD responses in SI and SII. Implicit mean image intensity normalization was not used during image processing.
A contrast comparing all blocks of stimulus ON with all blocks of stimulus OFF (baseline) was built for each individual experiment and for group analysis. T maps were coregistered with the spin-echo images of the whole brain and then superimposed for presentation. In some cases, line drawings from a rat brain atlas (Paxinos and Watson 1982) have been overlaid onto the images to delineate different cortical regions approximately. The cortical representations of whiskers (Fig. 1C) were used to define regions of interest in 5 contiguous slices centered around a slice positioned 2.5 mm caudal to bregma. We used the MarsBar tool (MARSeille Boîte À Région d'Intérêt) in SPM99 to measure the size of clusters, the changes in BOLD signal intensity, and P values for voxels in the regions of interest. A positive BOLD response or negative BOLD response was considered to be present in single-animal and group maps, if there was a cluster of 4 or more contiguous voxels that were all statistically significant (P < 0.05, uncorrected) in the region of interest (Forman et al. 1995). The volume of BOLD responses is reported as number of voxels. The volume in mm3 can be calculated by multiplying the number of voxels by 0.05445.
The cortical representation of rats' whiskers may show significant variability in mapping studies (Riddle and Purves 1995; Chen-Bee and Frostig 1996). This has direct ramifications on whether group map or single-subject analysis should be used (Woods 1996; Petersson et al. 1999a, 1999b; Thirion et al. 2007). Group maps derived from a fixed effects model are useful because they can reveal small-amplitude changes in the BOLD signal. In contrast, single-animal studies enable analysis of the variability that is known to occur between subjects. Therefore, group maps were used to identify changes in the BOLD signal and single-animal data were used to quantify those changes in more detail.
Statistics
Data are given as mean ± standard error of measurement (SEM) unless noted. We could not identify a suitable transform to stabilize the variance of the amplitudes of the positive BOLD response in SI at 3, 7, and 10 Hz. Therefore, we made pairwise comparisons of the 3, 7, and 10 Hz data and reduced the P value for statistical significance from 0.05 to 0.017 (Bonferroni correction for 3 possible pairwise comparisons).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingReferences |
|---|
BOLD Whisker Representations Depend on the Number of Deflected Whiskers
The response of neurons in somatosensory cortex to deflection of multiple whiskers is affected nonlinearly by several factors including the number of whiskers that contribute to the sensory input (Shimegi et al. 1999; Mirabella et al. 2001). It remains unclear, however, whether varying the number of deflected whiskers modifies the BOLD response. Therefore, we compared the BOLD responses evoked by deflection of the right C1 and C2 whiskers at 10 Hz with the BOLD responses evoked by 10-Hz deflection of multiple whiskers in the left B–E rows of the same animals (see Materials and Methods). Synchronous deflection of the right C1 and C2 whiskers evoked 2 positive BOLD responses in contralateral neocortex that were centered 2–3 mm caudal to bregma and extended over 2–5 contiguous slices of single-animal maps (Fig. 2A) and group maps (n = 8 rats; Fig. 3A). The location and rostrocaudal extent of the positive BOLD responses were consistent with previous neuroanatomical and electrophysiological studies of whisker cortical maps (Chapin and Lin 1984; Fabri and Burton 1991; Remple et al. 2003; Benison et al. 2007). We concluded that the positive BOLD responses represented activations in SI and secondary somatosensory cortex (SII). The positive BOLD responses were separated and partially surrounded by a negative BOLD response. In the group maps, negative BOLD responses were present in right-sided neocortex ipsilateral to the 2 deflected whiskers and in locations that were homotopic to the positive BOLD responses in contralateral SI and SII (Fig. 3Aii).
|
Deflection of 3–4 whiskers from each of the B–E whisker rows evoked a positive BOLD response in contralateral cortex that was centered 2–3 mm caudal to bregma and extended over 5–6 contiguous slices in single-animal (Fig. 2B) and group maps (Fig. 3B). The positive BOLD response was separated into 2 discrete responses in some single-animal maps (Fig. 2B) but not in the group map (Fig. 3B). In contrast to the 2-whisker protocol, there was no negative BOLD response contralateral to the deflected whiskers. Examination of the signal surrounding discrete positive BOLD responses in single-animal maps revealed that the BOLD signal was positive, but had not reached statistical significance. A negative BOLD response was present in somatosensory cortex ipsilateral to the deflected B–E whisker rows at points homotopic to the contralateral positive BOLD responses in both single-animal (–0.28 ± 0.06%, n = 16)(Fig. 2B) and group maps (Fig. 3B).
The peak amplitude of the positive BOLD response in SI evoked by deflecting 2 whiskers (+0.34 ± 0.04%) was less than that elicited by the multiple whisker–row protocol (+0.91 ± 0.08%; paired t-test, t = 6.74, n = 8, P < 0.001)(Fig. 4A). Similarly, the peak amplitude of the BOLD signal in SII evoked by 2-whisker deflection (+0.13 ± 0.05%) was smaller than that elicited by multiple whisker–row deflection (+0.84 ± 0.11%; paired t-test, t = 8.21, n = 8, P < 0.001) (Fig. 4B). The changes in volumes of the positive BOLD responses mirrored the differences in their amplitudes. The volume of the positive BOLD response in SI evoked by deflection of 2 whiskers (29 ± 5 voxels) was smaller than that elicited by the multiple whisker–row protocol (116 ± 18 voxels; paired t-test, t = 4.31, n = 8, P = 0.004) (Fig. 4C). Similarly, the volume of the positive BOLD response in SII evoked by deflection of 2 whiskers (16 ± 7 voxels) was less than that evoked by the multiple whisker–row protocol (66 ± 16 voxels; paired t-test, t = 2.62, n = 8, P = 0.034) (Fig. 4D). We concluded that the positive BOLD response evoked by deflection of 2 whiskers is of lower amplitude and is smaller in extent than the positive BOLD response evoked by deflection of many whiskers at 10 Hz.
|
Whisker Representations Are Frequency Dependent
The greatest decrease in deoxyhemoglobin levels and increase in CBV in the contralateral whisker barrel cortex evoked by deflection of one whisker occurs with 10-Hz stimulation (Sheth et al. 2003). Hence, BOLD responses evoked by repetitive deflection of 2 whiskers may vary across the physiological whisking range. We explored whether the cortical representation of the right C1 and C2 whiskers was frequency dependent by deflecting those whiskers synchronously: at 3 Hz, which is just below the physiological whisking range; at 7 Hz to represent exploratory whisking behavior; or at 10 Hz to exemplify discriminatory whisking behavior (Carvell and Simons 1995; Harvey et al. 2001). Deflection at 10 Hz evoked BOLD responses that were similar to those described earlier. Two positive BOLD responses were centered 2.5–3 mm caudal to bregma in contralateral neocortex and were abutted by a negative BOLD response in single-animal maps (Fig. 5) and group maps (Fig. 6C). In the group maps, 2 negative BOLD responses were present in right-sided neocortex ipsilateral to the whisker stimulation in locations that were homotopic to the positive BOLD responses in contralateral SI and SII (Fig. 6C). These data demonstrated that the positive and negative BOLD responses evoked by 10-Hz deflection of 2 whiskers were reproducible.
|
|
Whisker deflection at 3 Hz evoked a single positive BOLD response in a location consistent with SI (n = 7, Fig. 6A). Whisker stimulation at 7 Hz evoked a single positive BOLD response within contralateral neocortex that overlapped the 3-Hz activation but was larger in volume and extended laterally towards SII (n = 7; Fig. 6B). In contrast to the 10 Hz group map, no negative BOLD response was present in the 3-Hz or 7-Hz group maps contralateral to deflected whiskers. However, we noted that negative BOLD responses between SI and SII were elicited in single-animal maps (3/11 animals) with 7-Hz deflection when the positive BOLD response amplitude was large. Negative BOLD responses between contralateral SI and SII were infrequent in 3-Hz single-animal maps (1/11 animals). This suggested that the mechanisms driving the negative BOLD response during 10-Hz stimulation were also operating with 7-Hz stimulation, but more weakly. A negative BOLD response ipsilateral to the deflected whiskers was evoked in 45% (5/11) rats with whisker deflection at either 7 Hz (–0.28 ± 0.11%, n = 11) or 3 Hz (–0.14 ± 0.06%, n = 11).
We quantified the BOLD responses using regions of interest in the single-animal maps (Fig. 7A) and incorporated the 2-whisker data from the study comparing the representations of 2 whiskers with multiple whisker rows. Whisker deflection at 7 (n = 11) or 10 Hz (n = 16) always evoked a positive BOLD response in SI contralateral to the deflected whiskers. Deflection at 3 Hz elicited a contralateral positive BOLD response in the majority (9/11) of animals. The volume of the positive BOLD response in SI was correlated with the amplitude of the positive BOLD response in SI (r = 0.59, n = 38, P < 0.001) (Fig. 7B). Deflection frequency had a significant effect on the peak amplitude of the SI-positive BOLD response with larger responses at 7 Hz compared with either 3 Hz (P = 0.003, t = 3.4, t-test) or with 10 Hz (P = 0.008, T = 209, Mann–Whitney rank sum test) (3 Hz, 0.27 ± 0.05% [n = 11]; 7 Hz, 0.56 ± 0.07% [n = 11]; 10 Hz, 0.36 ± 0.02% [n = 16]; see Materials and Methods) (Fig. 7C). The volume of the positive BOLD response in SI of single animals was greater for 7-Hz whisker deflection compared with 3 Hz (P < 0.05, Dunn's multiple comparison) but did not attain statistical significance when 7 Hz was compared with 10-Hz deflection (3 Hz, 13 ± 4 voxels [n = 11]; 7 Hz, 41 ± 9 voxels [n = 11]; 10 Hz, 27 ± 3 voxels [n = 16]; P = 0.003, H = 11.8, Kruskal–Wallis 1-way ANOVA on ranks).
|
Positive BOLD responses were present in SII of 36% (4/11) of rats with 3-Hz stimulation, 54% (6/11) of rats with 7-Hz stimulation, and 50% (8/16) of rats with 10-Hz stimulation. The peak amplitude of the positive BOLD response in SII was not correlated with whisker-deflection frequency (3 Hz = 0.11 ± 0.05% [n = 11]; 7 Hz = 0.19 ± 0.06% [n = 11]; 10 Hz = 0.14 ± 0.04% [n = 16]: P = 0.507, H = 1.36, 1-way ANOVA on ranks) (Fig. 7C) in contrast to our results for SI. However, the amplitudes of BOLD responses in SII were close to the threshold for detection. Hence, the lack of correlation should be treated cautiously.
Our data indicated that the amplitude of the positive BOLD response in SI is greatest at a whisker-deflection frequency, 7 Hz, which is more typical of exploratory whisking behavior. Whisker deflection at higher frequencies to mimic discriminatory whisking behavior was associated with a decrease in amplitude of the positive BOLD response in SI and the emergence of a robust negative BOLD response that partially surrounded the positive BOLD responses in SI and SII (Figs 3A and 6C). We concluded that the cortical representation of 2 whiskers imaged with BOLD fMRI was frequency dependent in the physiological whisking range.
Positive BOLD Responses in Contralateral SI and SII
The thalamic inputs to SI and SII differ (Carvell and Simons 1987; Diamond 1995; Kwegyir-Afful and Keller 2004), which raises the possibility that BOLD signals in SI and SII may have distinct relationships with whisker-deflection frequency. However, reciprocal intracortical pathways connect SI and SII (Fabri and Burton 1991; Hoffer et al. 2003), and these pathways may couple BOLD responses in SI and SII. We found that the amplitude of the positive BOLD response was larger in SI compared with SII (P < 0.001, t = 6.7, n = 38, paired t-test) (Fig. 7D), but the amplitudes of the positive BOLD responses were not correlated (r = 0.18, P = 0.275, n = 38, Pearson correlation) (Fig. 7D). The lack of correlation (r = 0.28, P = 0.093, n = 37, Pearson correlation) persisted after removing one data point that was a possible outlier (3-Hz whisker deflection: SI, 0%; SII, 0.36%).
We counted the number of voxels in positive BOLD responses in SI and SII. The positive BOLD response in SI was spatially more extensive than the positive BOLD response in SII (positive BOLD response at 3, 7, and 10 Hz, median volume: SI = 22 voxels, SII = 5 voxels; n = 38, P < 0.001, signed rank test). The volumes of the positive BOLD responses in SI and SII were correlated (r = 0.46, P = 0.004, n = 38, Pearson correlation; Fig. 7E). We concluded that deflection of 2 adjacent whiskers evokes a positive BOLD response in SI and SII and that the positive BOLD response in SI is bigger than the positive BOLD response in SII. The lack of correlation between the amplitudes of the positive BOLD responses in SI and SII should be interpreted cautiously because the amplitudes of positive BOLD responses in SII were close to the threshold (approximately 0.15–0.2% signal change) for detection of responses in single animals. However, the results tend to suggest that BOLD responses in SI and SII are not extremely tightly coupled.
Positive and Negative BOLD Responses Have Different Laminar Locations
Inspection of the group maps of BOLD responses evoked by 10-Hz deflection of 2 whiskers (Figs 3A and 6C) suggested that the positive and negative BOLD responses were located at different depths in the cortex. We quantified this by dividing the SI and N regions of interest in whisker somatosensory cortex (Fig. 7A) into an upper half, which corresponds approximately to layers 1–4 (L1–4) and a lower half, which consists of layers 5–6 (L5–6) (Fig. 8A) and measured the volume of positive and negative BOLD responses in both halves of contralateral somatosensory cortex that were evoked by 10-Hz whisker deflection. The majority of the SI-positive BOLD response was in the upper cortical layers (L1–4 = 31 ± 6 voxels, L5–6 = 7 ± 3 voxels; P = 0.004, n = 16, t = 3.4, paired t-test), whereas the largest part of the negative BOLD response was in the lower cortical layers (L1–4 = 10 ± 3 voxels, L5–6 = 28 ± 7 voxels; P = 0.031, n = 16, t = 2.4, paired t-test) (Fig. 8B). We concluded that the BOLD response showed laminar preference with the positive BOLD response mainly located in upper cortical layers, whereas the negative BOLD response was predominantly found in lower cortical layers.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingReferences |
|---|
We imaged the steady-state cortical representations evoked by repetitive deflection of whiskers at multiple frequencies. Whisker representations depended on the number of displaced whiskers. Deflection of many whiskers evoked a positive BOLD response extending through contralateral SI and SII. In contrast, deflection of 2 adjacent whiskers at 10 Hz elicited positive BOLD responses in contralateral SI and SII, which were separated and partially surrounded by a negative BOLD response. The positive and negative BOLD responses showed different laminar specificities; positive BOLD responses were evoked predominantly in cortical layers 1–4, whereas negative BOLD responses were mainly found in cortical layers 5–6. Varying the whisker-deflection frequency in the physiological whisking range modified the amplitude and extent of the evoked BOLD responses. Our findings suggest that alterations in sensory input attributable to whisking movements modify whisker cortical representations.
Positive BOLD Responses
Deflection of many whiskers (multirow protocol) at 10 Hz evoked an extensive positive BOLD response that spread across contralateral SI and contralateral SII as reported previously (Sachdev et al. 2003; Lu et al. 2004; Kennerley et al. 2005). Similarly, focal stimulation of the whisker sensory input by deflecting 2 adjacent whiskers at 7 Hz resembled the positive BOLD response evoked by 8-Hz deflection of one whisker (Yang et al. 1996). In contrast, 2-whisker deflection at 10 Hz evoked very different representations with positive BOLD responses in SI and SII that were separated and partially surrounded by a negative BOLD response.
The limited resolution attainable with the BOLD signal (Ugurbil et al. 2003) might suggest that the separate positive BOLD responses in SI and SII with 10-Hz deflection were artifactual. It has proven difficult to map representations at the level of cortical columns using the BOLD signal and single-condition maps (Kim et al. 2000). However, refinements in functional imaging strategies such as differential mapping using the BOLD signal (Kim et al. 2000) or cerebral blood flow fMRI (Duong et al. 2001) have generated submillimeter resolution maps. Our ability to resolve separate positive BOLD responses in SI and SII was improved by the interspaced negative BOLD response. Furthermore, postprocessing of our data facilitated imaging of steady-state whisker representations. In particular, a vascular mask was used to minimize BOLD signal attributable to large draining veins and vascular inflow, which could markedly distort the location and extent of whisker representations (Menon and Goodyear 2001). A similar approach has been implemented to remove signal from surface vessels in optical imaging studies of single-whisker responses (Sheth et al. 2004a). Hence, we contend that the positive BOLD responses in SI and SII primarily represent separate foci of cortical activity evoked by whisker deflection.
The amplitude of the positive BOLD response evoked by deflection of many whiskers was larger than that elicited by 2-whisker deflection. We believe that this probably reflects a combination of greater thalamocortical input during multiple row–whisker deflection and spread of neuronal activity between barrel columns via horizontal intracortical connections (Moore and Nelson 1998; Zhu and Connors 1999; Brecht et al. 2003; Petersen et al. 2003).
Negative BOLD Responses
We imaged a robust negative BOLD response adjacent to positive BOLD responses in contralateral SI and SII with 2-whisker deflection at 10 Hz. This negative BOLD response is unlikely to be artifactual because it was not evoked by deflection of multiple whisker rows in the same animals despite the larger amplitude of the evoked positive BOLD response; our analysis suggests that it was frequency dependent; and it was not seen in scalp muscle near somatosensory cortex or in phantoms.
Negative BOLD responses adjacent to positive BOLD responses have been described in visual cortex of cats (Harel et al. 2002), humans (Shmuel et al. 2002), and monkeys (Shmuel et al. 2006) but have not been reported previously in rodent somatosensory cortex to our knowledge. The negative BOLD response that we found typically extends only 1–2 mm from the positive BOLD response, which is much less than that described for visual cortex (Harel et al. 2002; Shmuel et al. 2006). Our data further indicate that the negative BOLD response in somatosensory cortex evoked by whisker deflection is more prominent in lower cortical layers and is frequency dependent.
The mechanisms that underpin negative BOLD responses have attracted considerable debate with suggestions that they result from: diminution of neuronal activity (Raichle 1998); inhibitory synaptic activity (Blankenburg et al. 2003; Shmuel et al. 2006); neural activity insufficient to drive a vascular response (Nielsen and Lauritzen 2001; Sheth et al. 2004b); or a reduction in CBV (Harel et al. 2002) that may be caused by vascular steal (Kannurpatti and Biswal 2004). Furthermore, there has been discussion about the mechanisms that couple the negative BOLD response to neural activity and the role played by neurotransmitters released from inhibitory interneurons in neurovascular coupling (Attwell and Iadecola 2002; Shulman et al. 2004).
Negative BOLD responses were evoked in ipsilateral somatosensory cortex by both 2-whisker–row and multiple whisker–row deflection protocols. Performance of unilateral motor or sensory tasks in humans also evokes a negative BOLD response in ipsilateral sensorimotor cortex, which has been attributed to inhibition (Allison et al. 2000; Stefanovic et al. 2004). Such a mechanism could apply to the ipsilateral negative BOLD response that we describe because transcallosal connections linking right and left SI in the rat mediate feed-forward inhibition (Pidoux and Verley 1979). The lack of positive BOLD response in ipsilateral cortex suggests that 2-whisker deflection did not induce a significant increase in blood flow to ipsilateral cortex and, hence, implies that vascular steal is an unlikely cause.
Inhibition may underlie the contralateral negative BOLD response that we report. This proposal is supported by electrophysiological recordings, which indicate that synchronous deflection of 2 whiskers enhances surround inhibition (Simons and Carvell 1989; Brumberg et al. 1996) in addition to increasing neuronal activity in supragranular cortex (Shimegi et al. 1999; Mirabella et al. 2001). Furthermore, repetitive deflection in the upper part of the physiological whisking range and larger amplitude deflections entrain putative inhibitory interneurons better than excitatory neurons (Simons 1978). Whisker representations can extend outside whisker barrel cortex (Brett-Green et al. 2001). However, surround inhibition within whisker barrel cortex has a spatial gradient (Brumberg et al. 1996), and it is possible that the underlying inhibitory circuitry could result in the effects of surround inhibition being greater within whisker barrel cortex than in adjacent nonwhisker barrel cortex. Hence, 2-whisker deflection at 10 Hz may evoke negative BOLD responses because the protocol enables the effects of surround inhibition within whisker barrel cortex to be imaged. This proposal would explain why deflection of many whiskers did not evoke a contralateral negative BOLD response.
Non-fMRI imaging modalities have shown that somatosensory stimulation evokes an area of increased neuronal activity, increased oxygenation, and vasodilation in contralateral SI that is surrounded by a region of vasoconstriction and decreased oxygenation (Woolsey et al. 1996; Devor et al. 2005, 2007). Neurons in the surround region are hyperpolarized, which has been attributed to inhibition, but firing is unaffected (Kleinfeld and Delaney 1996; Derdikman et al. 2003; Devor et al. 2007). In contrast, negative BOLD responses in visual cortex are associated with both decreased cerebral blood flow (Harel et al. 2002; Shmuel et al. 2002) and diminished neuronal firing (Shmuel et al. 2006). The amplitude of the negative BOLD response correlates with the decrease in neuronal firing suggesting that diminished neural firing is an important factor underpinning negative BOLD responses (Shmuel et al. 2006).
We make a simple proposal bringing together hypotheses concerning the mechanisms underpinning negative BOLD. This synthesis centers on the balance between excitation and inhibition. A large proportion of excitatory synaptic inputs in neocortex arise from neighboring pyramidal neurons (Binzegger et al. 2004). Excitatory neurotransmitters released by those circuits cause vasodilation (Zonta et al. 2003; Takano et al. 2006; Wang et al. 2006). Therefore, inhibition that is strong enough to dampen firing has 2 effects, which both promote vasoconstriction. First, inhibition can cause vasoconstriction, possibly through a direct effect on vessels (Cauli et al. 2004). Second, diminished neural firing results in decreased release of excitatory neurotransmitters from local excitatory circuits and, hence, reduced vasodilatory drive. The combination of inhibition-induced vasoconstriction and reduced vasodilatory drive enhances the reduction in cerebral blood flow. A negative BOLD response would then occur in cortex, if the decrease in cerebral blood flow is much greater than the corresponding decrease in oxygen consumption.
Whisker Representations Are Frequency Dependent
The amplitude of the positive BOLD response evoked by 2-whisker deflection was frequency dependent with a peak response at 7 Hz. The mechanisms underpinning the frequency dependence of the positive BOLD response remain unclear. Optical imaging combined with local field potential measurements suggest that single-whisker responses are greatest with 10-Hz deflection (Sheth et al. 2003). In contrast, deflection of multiple whiskers evokes maximal synaptic responses (measured as the sum of local field potentials) with air puffs delivered at 5 Hz, and this correlates with increased astrocytic calcium levels (Wang et al. 2006), which cause vasodilatation (Takano et al. 2006). Preliminary reports of the BOLD response evoked by deflection of multiple whiskers have yielded mixed results (Lu et al. 2003; Melzer and Ebner 2004). The discrepancies in these studies may be due to configuration of the whisker stimulus, which markedly affects evoked responses (Pinto et al. 2000; Petersen et al. 2003).
The emergence of a robust negative BOLD response in the 10 Hz group maps elicited by 2-whisker deflection suggested that the boundary of the positive BOLD response was modulated by whisker deflection in the upper part of the physiological whisking range. Changes in representation boundaries, referred to as sharpening, have been reported with similar whisker-stimulation frequencies (Sheth et al. 1998). The whisker deflections that we used were designed to mimic whisking and, therefore, may have had larger amplitudes. However, the boundary changes that we imaged with 10-Hz deflection of 2 whiskers resemble sharpening of whisker representations, suggesting the possibility that the contralateral negative BOLD and sharpening of whisker representations are driven by similar mechanisms.
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingReferences |
|---|
Wellcome Trust and MRC[P4].
|
Acknowledgments |
|---|
We thank Professor Mick Brammer and Claire Cheetham for helpful comments and Professor Jimmy Bell, Dr Amy Herlihy, and Dr Po-Wah So for support with the project. All scans were performed at the MRC Biological Imaging Centre, Imperial College London. GTF is a Wellcome Trust Senior Fellow in the Clinical Sciences. Conflict of Interest: None declared.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingReferences |
|---|
Allison JD, Meador KJ, Loring DW, Figueroa RE, Wright JC. Functional MRI cerebral activation and deactivation during finger movement. Neurology. (2000) 54:135–142.
Attwell D, Iadecola C. The neural basis of functional brain imaging signals. Trends Neurosci. (2002) 25:621–625.[CrossRef][ISI][Medline]
Austin VC, Blamire AM, Allers KA, Sharp T, Styles P, Matthews PM, Sibson NR. Confounding effects of anesthesia on functional activation in rodent brain: a study of halothane and alpha-chloralose anesthesia. Neuroimage. (2005) 24:92–100.[CrossRef][ISI][Medline]
Benison AM, Rector DM, Barth DS. Hemispheric mapping of secondary somatosensory cortex in the rat. J Neurophysiology. (2007) 97:200–207.
Binzegger T, Douglas RJ, Martin KA. A quantitative map of the circuit of cat primary visual cortex. J Neurosci. (2004) 24:8441–8453.
Blankenburg F, Taskin B, Ruben J, Moosmann M, Ritter P, Curio G, Villringer A. Imperceptible stimuli and sensory processing impediment. Science. (2003) 299:1864.
Brecht M, Roth A, Sakmann B. Dynamic receptive fields of reconstructed pyramidal cells in layers 3 and 2 of rat somatosensory barrel cortex. J Physiol (Lond). (2003) 553:243–265.
Brett-Green BA, Chen-Bee CH, Frostig RD. Comparing the functional representations of central and border whiskers in rat primary somatosensory cortex. J Neurosci. (2001) 21:9944–9954.
Brumberg JC, Pinto DJ, Simons DJ. Spatial gradients and inhibitory summation in the rat whisker barrel system. J Neurophysiology. (1996) 76:130–140.
Carvell GE, Simons DJ. Thalamic and corticocortical connections of the second somatic sensory area of the mouse. J Comp Neurol. (1987) 265:409–427.[CrossRef][ISI][Medline]
Carvell GE, Simons DJ. Biometric analyses of vibrissal tactile discrimination in the rat. J Neurosci. (1990) 10:2638–2648.[Abstract]
Carvell GE, Simons DJ. Task- and subject-related differences in sensorimotor behavior during active touch. Somatosens Mot Res. (1995) 12:1–9.[ISI][Medline]
Castro-Alamancos MA. Absence of rapid sensory adaptation in neocortex during information processing states. Neuron. (2004) 41:455–464.[CrossRef][ISI][Medline]
Cauli B, Tong XK, Rancillac A, Serluca N, Lambolez B, Rossier J, Hamel E. Cortical GABA interneurons in neurovascular coupling: relays for subcortical vasoactive pathways. J Neurosci. (2004) 24:8940–8949.
Chapin JK, Lin C-S. Mapping the body representation in the SI cortex of anesthetized and awake rats. J Comp Neurol. (1984) 229:199–213.[CrossRef][ISI][Medline]
Chen-Bee CH, Frostig RD. Variability and interhemispheric asymmetry of single-whisker functional representations in rat barrel cortex. J Neurophysiology. (1996) 76:884–894.
Crochet S, Petersen CCH. Correlating whisker behavior with membrane potential in barrel cortex of awake mice. Nat Neurosci. (2006) 9:608–610.[CrossRef][ISI][Medline]
Derdikman D, Hildesheim R, Ahissar E, Arieli A, Grinvald A. Imaging spatiotemporal dynamics of surround inhibition in the barrels somatosensory cortex. J Neurosci. (2003) 23:3100–3105.
Derdikman D, Yu C, Haidarliu S, Bagdasarian K, Arieli A, Ahissar E. Layer-specific touch-dependent facilitation and depression in the somatosensory cortex during active whisking. J Neurosci. (2006) 26:9538–9547.
Devor A, Tian P, Nishimura N, Teng IC, Hillman EM, Narayanan SN, Ulbert I, Boas DA, Kleinfeld D, Dale AM. Suppressed neuronal activity and concurrent arteriolar vasoconstriction may explain negative blood oxygenation level-dependent signal. J Neurosci. (2007) 27:4452–4459.
Devor A, Ulbert I, Dunn AK, Narayanan SN, Jones SR, Andermann ML, Boas DA, Dale AM. Coupling of the cortical hemodynamic response to cortical and thalamic neuronal activity. Proc Natl Acad Sci USA. (2005) 102:3822–3827.
Diamond ME. Somatosensory thalamus of the rat. In: The barrel cortex of rodents—Jones EG, Diamond IT, eds. (1995) New York: Plenum Press. 189–220.
Duong TQ, Kim DS, Ugurbil K, Kim SG. Localized cerebral blood flow response at submillimeter columnar resolution. Proc Natl Acad Sci USA. (2001) 98:10904–10909.
Fabri M, Burton H. Ipsilateral cortical connections of primary somatic sensory cortex in rats. J Comp Neurol. (1991) 311:405–424.[CrossRef][ISI][Medline]
Fanselow EE, Nicolelis MA. Behavioral modulation of tactile responses in the rat somatosensory system. J Neurosci. (1999) 19:7603–7616.
Fee MS, Mitra PP, Kleinfeld D. Central versus peripheral determinants of patterned spike activity in rat vibrissa cortex during whisking. J Neurophysiology. (1997) 78:1144–1149.
Ferezou I, Bolea S, Petersen CC. Visualizing the cortical representation of whisker touch: voltage-sensitive dye imaging in freely moving mice. Neuron. (2006) 50:617–629.[CrossRef][ISI][Medline]
Forman SD, Cohen JD, Fitzgerald M, Eddy WF, Mintun MA, Noll DC. Improved assessment of significant activation in functional magnetic resonance imaging (fMRI)—use of a cluster size threshold. Magn Reson Med. (1995) 33:636–647.[ISI][Medline]
Friston KJ, Williams S, Howard R, Frackowiak RS, Turner R. Movement-related effects in fMRI time-series. Magn Reson Med. (1996) 35:346–355.[ISI][Medline]
Harel N, Lee SP, Nagaoka T, Kim DS, Kim SG. Origin of negative blood oxygenation level-dependent fMRI signals. J Cereb Blood Flow Metab. (2002) 22:908–917.[CrossRef][ISI][Medline]
Harvey MA, Bermejo R, Zeigler HP. Discriminative whisking in the head-fixed rat: optoelectronic monitoring during tactile detection and discrimination tasks. Somatosens Mot Res. (2001) 18:211–222.[CrossRef][ISI][Medline]
Hentschke H, Haiss F, Schwarz C. Central signals rapidly switch tactile processing in rat barrel cortex during whisker movements. Cereb Cortex. (2006) 16:1142–1156.
Hlustik P, Noll DC, Small SL. Suppression of vascular artifacts in functional magnetic resonance images using MR angiograms. Neuroimage. (1998) 7:224–231.[CrossRef][ISI][Medline]
Hoffer ZS, Hoover JE, Alloway KD. Sensorimotor corticocortical projections from rat barrel cortex have an anisotropic organization that facilitates integration of inputs from whiskers in the same row. J Comp Neurol. (2003) 466:525–544.[CrossRef][ISI][Medline]
Kannurpatti SS, Biswal BB. Negative functional response to sensory stimulation and its origins. J Cereb Blood Flow Metab. (2004) 24:703–712.[ISI][Medline]
Kennerley AJ, Berwick J, Martindale J, Johnston D, Papadakis N, Mayhew JE. Concurrent fMRI and optical measures for the investigation of the hemodynamic response function. Magn Reson Med. (2005) 54:354–365.[CrossRef][ISI][Medline]
Kim D-S, Duong TQ, Kim S-G. High-resolution mapping of iso-orientation columns by fMRI. Nat Neurosci. (2000) 3:164–169.[CrossRef][ISI][Medline]
Kleinfeld D, Delaney KR. Distributed representation of vibrissa movement in the upper layers of somatosensory cortex revealed with voltage-sensitive dyes. J Comp Neurol. (1996) 375:89–108.[CrossRef][ISI][Medline]
Kwegyir-Afful EE, Keller A. Response properties of whisker-related neurons in rat second somatosensory cortex. J Neurophysiology. (2004) 92:2083–2092.
Kwong KK, Belliveau JW, Chesler DA, Goldberg IE, Weisskoff RM, Poncelet BP, Kennedy DN, Hoppel BE, Cohen MS, Turner R, et al. Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation. Proc Natl Acad Sci USA. (1992) 89:5675–5679.
Lowe AS, Barker GJ, Beech JS, Ireland MD, Williams SC, Forthcoming. A method for removing global effects in small animal fMRI. NMR Biomed. (2007).