Cerebral Cortex Advance Access originally published online on October 5, 2007
Cerebral Cortex 2008 18(6):1361-1373; doi:10.1093/cercor/bhm168
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Plasticity of Representational Maps in Somatosensory Cortex Observed by In Vivo Voltage-Sensitive Dye Imaging
Department of Cell Physiology, Max Planck Institute for Medical Research, Jahnstrasse 29, D-69120 Heidelberg, Germany
Address correspondence to Damian J. Wallace, Department of Cell Physiology, Max Planck Institute for Medical Research, Jahnstrasse 29, D-69120 Heidelberg, Germany. Email: dhw{at}mpimf-heidelberg.mpg.de.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionConclusionsFundingReferences |
|---|
We investigated the effect of selective whisker trimming on the development of the cortical representation of a whisker deflection in layer 2/3 of rat somatosensory cortex using in vivo voltage-sensitive dye (vsd) imaging. Responses to deflection of D-row whiskers were recorded after trimming of A-row, B-row, and C-row whiskers, referred to as DE pairing, during postnatal development. Animals DE paired from postnatal day (p) 7 to p17 had a significant bias in the spread of the vsd signal, favoring spread toward the concomitantly nondeprived E-row columns. This resulted primarily from a strong decrease in signal spreading into the deprived C-row columns. In contrast, signal spread in control littermates was approximately symmetrical. DE pairing failed to elicit significant changes when begun after p14, thus defining a critical period for this phenomenon. The results suggest that sensory deprivation in this model results in lower connectivity being established between nondeprived columns and adjacent deprived ones.
Key Words: barrel cortex • cortical plasticity • critical period • sensory deprivation • whisker
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsFundingReferences |
|---|
The capacity to adapt to changes in incoming sensory information is a hallmark of the mammalian cerebral cortex. Such "cortical plasticity" is observed in different cortical areas, can act to increase or decrease cortical responsiveness, and is differentially expressed in different cortical layers (Weinberger 2004
; Feldman and Brecht 2005; Hensch 2005; Ohl and Scheich 2005; Sur and Rubenstein 2005). The rodent barrel cortex, the somatosensory area representing the facial whiskers on the animals' snout, is a convenient model for studying mechanisms generating this cortical plasticity for several reasons. First, barrel cortex is highly plastic from birth into adulthood, displaying several different types of usage-dependent change in cortical function (Simons and Land 1987; Fox 1992; Glazewski and Fox 1996; Glazewski et al. 1998; Polley et al. 1999; Feldman and Brecht 2005). Second, use-dependent plasticity can be induced by trimming or removing whiskers, a simple and noninvasive manipulation (Simons and Land 1987; Fox 1992). Finally, the precisely ordered topography of the whisker representation greatly facilitates identification of deprived and spared cortical columns both during and after recordings of responses to sensory stimulation.
Different types of map plasticity are expressed in barrel cortex during different stages of the animals' life and differentially in the different cortical layers. Map changes have been reported for granular, supragranular, and infragranular layers (Simons and Land 1987; Fox 1992; Diamond et al. 1993, 1994), though plasticity in layer 4, presumably reflecting changes in the thalamocortical projection, occurs more reluctantly than in the other layers, requiring either trimming during a short critical period during the first postnatal week (Fox 1992), relatively long periods of whisker trimming (Armstrong-James et al. 1994), or a "chessboard" pattern of whisker trimming (Wallace and Fox 1999). In layer 2/3, whisker deprivation has been shown to result in reductions in the responsiveness of cells within deprived cortical columns (Glazewski et al. 1998) as well as increasing responsiveness in the cortical columns representing spared (i.e., untrimmed) whiskers (Glazewski and Fox 1996). In rats reared in standard laboratory cages, trimming all but one whisker ("single-whisker sparing") results in increased responsiveness of layer 2/3 cells in the column representing the nontrimmed whisker and in an increase in the extent of the cortical representation of the nontrimmed whisker at the expense of the representations of the surrounding trimmed whiskers (Fox 1992; Polley et al. 1999). Intriguingly, the reverse effect, that is a reduction in the extent of the cortical representation of the spared whisker, is observed in adult rats trimmed of all but one whisker on one side of the face and subsequently placed for brief periods in an environment encouraging whisker-dependent exploration (Polley et al. 1999). It has also been shown that sparing of 2 neighboring D-row whiskers results in mutually enhanced responsiveness between the 2 spared whiskers without expansion into the surrounding deprived regions (Diamond et al. 1993; Armstrong-James et al. 1994; Diamond et al. 1994; Lebedev et al. 2000). All the above observations indicate the initiation of some form of alterations to the underlying neuronal circuits and subsequent alteration to the transmission of excitation between used and unused cortical areas. However, it is currently unclear what underlies these changes, for example, whether the changes are represented at the level of both sub- and suprathreshold activity.
Recent studies employing voltage-sensitive dye (vsd) imaging in vivo have shown this technique to be well suited for visualizing the spread of activity between cortical columns (Kleinfeld and Delaney 1996; Derdikman et al. 2003; Petersen et al. 2003). Vsd studies in barrel cortex show that whisker deflections evoke a vsd response which is initially restricted to an area about the size of the column of the stimulated whisker and that the signal then rapidly spreads into the surrounding columns (Kleinfeld and Delaney 1996; Petersen et al. 2003). The method used for in vivo vsd staining typically results in the dye penetrating and labeling tissue to a depth of 900–1000 µm, thus the signals are thought to predominantly reflect the electrical activity of layer 2/3 (Kleinfeld and Delaney 1996; Petersen et al. 2003). Furthermore, simultaneous intracellular recording and vsd imaging indicate that the vsd signals in vivo provide a reliable measure of subthreshold changes in membrane potential in the stained cortical area (Petersen et al. 2003). Thus, vsd imaging is a convenient experimental approach for examining usage-dependent effects on the spread of synaptic activity across the cortex in layer 2/3.
Here, we have used in vivo vsd imaging to examine the effects of sensory deprivation on the spread of subthreshold activity from the stimulated principal whisker (PW) into surround whisker (SuW) columns. The deprivation paradigm we have used is an extension of the whisker-pairing paradigm initially introduced by Diamond et al. (1993) and involves pairing the D- and E-rows of whiskers (DE pairing), by trimming the A-, B-, and C-rows. This paradigm surrounds the D-row columns with "used" cortical columns, that is, columns with preserved whisker-driven sensory input, on one side and deprived or "unused" areas on the other. Finnerty et al. (1999) previously used the same deprivation protocol to investigate its effects on the short-term dynamics the relevant intracortical pathways. Based on the conclusions of these previous studies, we predict that the DE-pairing paradigm should result in an increase in the effective connectivity between adjacent spared columns. With reference to the changes across the border between spared and deprived columns, the above studies differ in their results. The conclusions of the study by Finnerty et al. predict an increase in effective connectivity, whereas those of Diamond et al. would predict no substantial change. We conducted vsd-imaging experiments to determine whether the DE-pairing paradigm resulted in changes in the spatial dynamics of the vsd signal following a brief deflection of a D-row whisker.
Whisker trimming has been shown to have an effect on the normal development of sensory receptive fields of layer 2/3 neurons (Stern et al. 2001), an effect with a critical period ending around postnatal day (p) 14. It is approximately during this developmental period that the axons of layer 2/3 pyramidal neurons are extending at their most prolific rate (Radnikow G, Feldmeyer D, personal communication) and the layer 2/3 neurons are developing mature input connections from layer 4 (Stern et al. 2001; Bureau et al. 2004). We therefore also investigated whether this deprivation paradigm had the same influence on patterns of cortical connectivity if animals were deprived during this critical period of development of the layer 2/3 network compared with later in development after this network has been more fully established.
In brief, we find that DE pairing during the second and third postnatal week results in a significant reduction in the spread of the vsd signal into cortical columns representing the trimmed whiskers. The critical period for this effect ends at the end of the second postnatal week, and the effect is reversed within 10 days of restoration of afferent activity to the previously deprived cortical areas.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsFundingReferences |
|---|
Whisker Deprivation
All animal procedures were carried out according to the animal handling regulations of the Max Planck Society. Whisker trimming was accomplished by daily clipping of the appropriate rows of whiskers to the level of the facial fur. Animals used as age- and litter-matched controls for the deprived animals were handled in the same way and for approximately the same amount of time as whisker-trimmed animals, and all animals were housed with their mothers with free access to food and water until used for experimental recordings. The pattern and duration of whisker trimming and age of animals at the beginning of the deprivation period varied between experimental groups and is stated in the text prior to the description of results.
Surgery and Vsd Imaging
On the day of experimental recordings, animals were anaesthetized by intraperitoneal injection of urethane (1.6 g/kg body weight). Supplementary doses of urethane (20% of original dose) were administered as required to maintain anesthetic depth at a level where paw withdrawal and corneal reflexes were absent. Body temperature was monitored throughout the experiment with a rectal probe and was maintained at approximately 37 °C.
The skull was exposed and a craniotomy measuring approximately 4 x 4 mm opened over the left barrel field centered approximately at 2.5 mm posterior and 5.5 mm lateral to bregma. The dura was removed and edema minimized by draining the cisternal magna. The cortex was then stained with the vsd RH1691 (1 mg/ml; Optical Imaging, Ltd, Rehovot, Israel) for a period of approximately 120 min by direct topical application of the dye onto the cortical surface. The RH1691 was dissolved in Ringer's solution containing 135 mM NaCl, 5 mM KCl, 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, 1.8 mM CaCl2, and 1 mM MgCl2. After the staining period, the craniotomy was washed with Ringer's solution to remove unbound dye, covered with a 1% agar solution, and capped with a coverslip for imaging. Images of vsd responses were acquired using equipment and procedures as described in detail previously (Petersen et al. 2003). Briefly, images were acquired using a Fuji HR Deltaron 1700 camera (Fuji, Tokyo, Japan) modified according to Shoham et al. (1999) and tandem lens macroscope as described by Ratzlaff and Grinvald (1991). The focal plane of the macroscope was set to approximately 350 µm below the pia, and frames were collected every 4.8 ms. Excitation light was filtered through a 630 ± 30-nm band-pass filter, and reflected responses were collected after filtering through a 665-nm long-pass filter. Acquisition was triggered on the electrocardiogram, and sweeps were collected both with and without whisker stimulus for correction of bleaching and heart beat artifacts by subtraction. Averages of 30–50 collected stimulated sweeps were made in order to minimize the contribution of ongoing spontaneous activity on the records and to obtain a reliable estimation of the average evoked cortical response.
Whisker Stimulation
Individual whiskers were stimulated using a computer-controlled piezoelectric wafer with a 500 µm caudal deflection lasting 2 ms and delivered at 10 mm from the whisker pad, amounting to an approximately 3 degree deflection of the whisker at the follicle.
Data Analysis
Data presented in figures have been low-pass filtered for clarity; however, all analyses used raw data and were conducted using custom-written routines executed in Matlab software (Mathworks, MA). Initial image processing consisted of normalization to the reflected light intensity. This was achieved by calculating a "normalization frame", that being a pixel-wise average of all frames in the 180-ms period prior to the stimulus, and subsequently performing a pixel-wise division of all recorded frames in the sweep by the normalization frame. This was followed by a "blank subtraction" to remove bleaching and heart beat artifacts. This blank subtraction involved calculating a frame-wise average of all sweeps without stimulus and a corresponding frame-wise average of all sweeps with stimulus and finally performing a frame-wise subtraction of the nonstimulated frames from the stimulated frames. This then resulted in the final 
F/F0 values for each pixel in each frame.
To analyze the characteristics of the vsd signal as it spreads from the stimulated whisker column into surrounding columns, an arbitrary row axis was assigned for each individual whisker response based on functional responses obtained from stimulation of 2 neighboring in row whiskers (Supplementary Fig. S1). For each of these responses the center of the stimulated barrel was defined as the pixel with the greatest intensity in a low-pass filtered version of the earliest poststimulus image showing a clear response and the arbitrary row axis defined as a straight line fit between the 2 barrel centers.
To examine the symmetry of responses within the stimulated barrel, 2 semicircular regions of interest (ROIs) with a radius of 200 µm and oriented parallel with the row axis were defined. The average response intensity within the 2 ROIs was then calculated for all acquired image frames.
Responses spreading into first-order SuW columns were calculated as the average response intensity within 2 arc-shaped ROIs. The inner arc of these ROIs had a radius of 400 µm from the center of the stimulated whisker column and the outer arc a radius of 600 µm. The arcs started at an angle of 45 degrees on either side of the row axis and subtended an angle of 90 degrees, and responses within the ROIs were again calculated for all acquired image frames.
C-row to E-row ROI ratios were calculated by dividing the response quantified in the C-row ROI by that in the E-row ROI independently for each image frame within each whisker response.
Segment-to-PW ROI ratios were calculated by quantifying the response within the arc-shaped ROI described above and dividing this by the response quantified within the semicircular ROI described above independently for each image frame within each whisker response.
Three groups of animals were analyzed in the series of experiments examining the effects of DE pairing during the second and third postnatal week. One group had D- and E-rows of whiskers paired from p7 and p17, the second were DE-row paired for a shorter period of 5 days between p9 and p14, and the third group which were D2E2 paired (by trimming of all whiskers apart from D2 and E2) from p7 to p17. The effect of these protocols on the preferential spread of activity around stimulated D-row whiskers was not different, thus these groups have been pooled for the analysis of the effect of DE pairing during the second and third postnatal week.
All data reported and graphs shown are mean responses ± standard error of the mean. Statistical testing was with unpaired Student's t-tests unless otherwise stated.
Histology and Reconstruction of Barrel Patterns
At the termination of each experiment, the animal was perfused transcardially initially with phosphate buffer (pH 7.2) and subsequently with 4% paraformaldehyde (PFA) in phosphate buffer. The brain was then removed and postfixed in 4% PFA overnight at 4 °C. Vibratome sections (150 µm thick) were cut, and the barrel pattern in layer 4 of the somatosensory cortex was revealed by cytochrome oxidase staining (Wong-Riley 1979). For illustrations of the layer 4 barrel pattern in relation to the vsd responses, reconstructions of the barrels were made from these stained sections using Neurolucida software (MicroBrightField, Colchester, VT) and aligned to the vsd images using either the vascular pattern on the cortical surface or the early vsd frames (10–20 ms poststimulus) recorded from single-whisker stimulation of 2 or more whiskers.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsFundingReferences |
|---|
DE Pairing Results in Asymmetrical Spread of Vsd Responses
We measured changes in the spatiotemporal dynamics of cortical responses resulting from a cortical area being surrounded on one side by a used (nondeprived) cortical area and by an unused (deprived) area on the other. We chose to focus on manipulations of sensory input around D-row whiskers because it has first-order SuWs on all sides, and the whiskers are easily accessed for stimulation after the surgery to expose barrel cortex. All responses described are those evoked by deflection of a single D-row whisker.
Pulsatile whisker deflection evoked a response that could be first observed in vsd images in animals at this age (p18–p23 at the time of the experiment) at 10–16 ms poststimulus (Fig. 1A). The earliest detectable vsd response was spatially restricted to a small area approximately bounded by the extent of the borders of the layer 4 barrel of the stimulated whisker, as previously reported (Petersen et al. 2003). Within approximately 50 ms, the response had spread to cover a major fraction of the whisker representation in the primary somatosensory cortex (Fig. 1A). Consistent with the reported pattern of axonal projections for layer 2/3 neurons in barrel cortex (Bernardo et al. 1990), there was a tendency for the vsd responses to spread preferentially along rows compared with across arcs (Supplementary Fig. S2). DE pairing during the second and third postnatal week (p7–p17) resulted in an alteration in the time-dependent spread of the vsd responses across the barrel cortex. After DE pairing, vsd responses spread preferentially towards the cortical areas representing the spared E-row whiskers (Fig. 1B). Fifteen of the 16 D-row whisker responses recorded from DE-paired animals showed asymmetrical lateral spread of the vsd response. This effect on the shape of the area depolarized in the first 50 ms poststimulus is seen clearly by comparison of the isointensity contours in Figure 2.
|
|
Symmetry of Spread within the PW Column
As mentioned above, the vsd responses have a tendency to spread preferentially along the rows. However, the focus of the current study was to examine the effect of asymmetrical afferent cortical activity around D-row columns, that is, the effect of having unused cortical columns on one side and used columns on the other. In this context, for the remainder of the manuscript, we describe response symmetry in terms of the symmetry of the vsd response along the arc axis. That is, the symmetry of the response spreading toward the spared and the deprived whisker-associated columns.
We first measured the symmetry of vsd responses within the area of the stimulated PW column. We determined the average response intensity in 2 semicircular ROIs of radius 200 µm aligned along the axis of the columns representing D-row whiskers (Fig. 3A). As shown in Figure 3B1, responses quantified within these ROIs for control animals were almost perfectly overlapping (pooled data from 21 D-row whisker responses from 9 animals shown in Fig. 3C). Consequently, average C-row to E-row ROI ratios (CE ratios) calculated for these responses were very close to 1. DE pairing had no significant effect on the symmetry of responses within the principal column of the stimulated whisker (Fig. 3B2, pooled data shown in Fig. 3D are from 16 D-row whisker responses from 10 animals), with CE ratios again very close to 1. We interpret this as indicating that the symmetry of the net postsynaptic depolarization within the layer 2/3 network of the PW column evoked by feedforward excitation from layer 4 and by direct thalamocortical excitation is unaffected by DE pairing.
|
Lateral Spread across Rows
To quantitatively compare the symmetry of vsd responses as they spread laterally from the stimulated whiskers' column, we calculated the average response magnitude in 2 ROIs located over the neighboring first-order SuW areas (Fig. 4; see Materials and Methods for full details of ROI parameters). Responses in control animals were slightly asymmetrical, showing a tendency to spread preferentially toward neighboring C-row areas (Fig. 4B1, pooled data in Fig. 4C, and data set as above). Corresponding average CE ratios were thus slightly greater than 1 (Fig. 4E). After DE pairing, responses spread preferentially toward E-row areas (Fig. 4B2, pooled data in Fig. 4D). This reversal in the direction of preferential spread of activity was reflected in a substantial reduction in CE ratios to values well below 1 (Fig. 4E).
|
For statistical comparisons between the 2 groups of animals, we generated a single image of the average vsd response by calculating an average of all individual image frames from 20 to 50 ms poststimulus, referred to from here on as an Av20to50 image. Responses were quantified in the Av20to50 images within the same ROIs as described above. The reduction in the average CE ratio calculated in Av20to50 images was highly significant (control average 1.04 ± 0.02, 21 D-row whisker responses from 9 animals, DE-paired average 0.87 ± 0.02, 16 D-row whisker responses from 10 animals; P < 0.001).
We also tested whether the effects of the DE-pairing protocol were dependent on pairing rows of whiskers by examining the symmetry of responses in animals after D2E2 pairing (trimming of all whiskers apart from D2 and E2) from p7 to p17. Responses to stimulation of the spared D2 whisker showed very pronounced asymmetry in all animals tested. The average CE ratio calculated in Av20to50 images was also significantly different to the control group (D2E2-paired average CE ratio 0.82 ± 0.05, 4 D2 whisker responses from 4 animals; P < 0.01). The effect of the DE pairing is thus not dependent on the presence of whole rows of whiskers but can also be induced by pairing of 2 in-arc neighboring whiskers.
CD-Row Whisker Pairing
We conducted a further set of experiments to test the effect of pairing C- and D-rows of whiskers on preferential spread of cortical postsynaptic depolarization around the columns of stimulated D-row whiskers. In these experiments, CD pairing was maintained from p7 to p17 with the hypothesis that if the plasticity of the vsd responses we had observed after DE pairing were consistent in its action across the barrel field, we would observe an increase in the bias in spread of the vsd signals toward the used, undeprived C-row areas and consequently an increase in the observed CE ratios. Of 6 D-row whisker responses tested from 3 animals after CD pairing, 5 showed a strong bias in the spread of cortical activity towards C-row areas, whereas one response showed a bias instead toward E-row. The average CE ratio time course for responses after CD pairing is shown in Figure 4E. The change in average CE ratios calculated in Av20to50 images for this experimental series was, however, not statistically significant (Control 0.97 ± 0.08, 6 D-row whisker responses from 3 animals; CD paired 1.17 ± 0.10, 6 D-row whisker responses from 3 animals; P = 0.07).
Overall, these results suggest that pairing rows of whiskers during early postnatal development induces a bias in the spread of subthreshold activity that favors neighboring, active cortical areas.
Reduction in the Spread of Activity between Used and Deprived Cortical Areas
We next made an analysis aimed at establishing whether the observed changes in the spread of activity through the layer 2/3 network were the result of enhanced propagation between the 2 used cortical areas or a reduction in propagation between the used and the deprived cortical areas. This involved computing for each individual whisker response, the ratio of the response recorded in the first-order SuW segment ROI to that recorded in the PW ROI. The ratios of SuW segment-to-PW ROIs for each time point were averaged across all whisker responses to give the average temporal profiles shown in Figure 5. The analysis revealed that DE pairing resulted in a major reduction in the SuW segment-to-PW ratio calculated for C-row areas (Fig. 5A1) and concomitantly a more modest reduction in ratios calculated for E-row areas (Fig. 5A2). Consistent with this result, when the same analysis was done for the experiments using the CD-pairing protocol, SuW segment-to-PW ratios calculated for the deprived E-row areas were reduced, whereas ratios calculated for the paired C-row areas were unchanged (Fig. 5B1,B2).
|
For statistical comparison, we calculated the SuW segment-to-PW ratio in Av20to50 images. After DE pairing, the average SuW segment-to-PW ratio for C-row areas was significantly reduced in the DE-paired group (control 0.69 ± 0.03, deprived 0.53 ± 0.03, P < 0.001, data sets for control and DE-paired responses as described above). The average ratio for E-row areas was also reduced, though the reduction in this case was more modest (control 0.64 ± 0.02, deprived 0.58 ± 0.02, P = 0.07). For the CD-pairing experimental series, segment-to-PW ratios calculated for the deprived E-row areas were reduced after CD pairing, though the reduction was not statistically significant (control 0.65 ± 0.04, deprived 0.58 ± 0.02, P = 0.15, data sets for control and CD-paired responses as described above). SuW segment-to-PW ratios for the C-row areas were unchanged after CD pairing (control 0.64 ± 0.05, deprived 0.67 ± 0.03, P = 0.69).
Taken together, these results suggest that the alteration in the spread of the vsd signal after DE pairing was primarily due to a reduction in the propagation of subthreshold responses between the used and the deprived cortex.
Pairing Effect Has a Critical Period Ending around the End of the Second Postnatal Week
The above results show that pairing rows of whiskers during the second and third week of postnatal development introduce a bias in the spread of activity in layer 2/3, primarily by reducing the spread of activity between areas with active sensory input and those without. During this developmental period, layer 2/3 neurons are extending axonal arbors at their most rapid rate (Radnikow G, Feldmeyer D, personal communication), are developing their synaptic responses to whisker stimulation, and establishing their receptive field properties (Stern et al. 2001; Bureau et al. 2004). We were therefore interested in establishing whether the effect we have observed here also is observed in more mature cortex. To test this, a group of animals was DE paired from p21 to p31 and responses were tested after this period of whisker deprivation. Representative responses for a control and DE-paired animal are shown in Figure 6A. DE pairing during this later period had little effect on the spread of cortical activity around the stimulated whisker. Average responses quantified in surround segment ROIs showed a slight tendency toward increased response magnitude in the C-row ROIs after DE pairing (Fig. 6B compared with Fig. 6C); however, average CE ratios calculated from Av20to50 images were not significantly different (average C-row to E-row ROI ratio for control responses 1.02 ± 0.03, 19 D-row whisker responses from 7 control animals, average for DE-paired animals 1.06 ± 0.03, 21 D-row whisker responses from 7 DE-paired animals, P = 0.23).
|
As was found to be the case for the younger animals in the early postnatal row-pairing experimental groups, responses measured in the C-row and E-row PW ROIs were almost identical for both the control animals and the animals DE paired from p21 to p31 (Supplementary Fig. S3). Surround segment-to-PW ROI ratios calculated as outlined above were not different after DE pairing in these animals (Supplementary Fig. S4). This was also borne out in statistical comparisons of surround segment-to-PW ROI ratios from Av20to50 images (C-row control average 0.75 ± 0.02, C-row deprived average 0.77 ± 0.02, P = 0.42; E-row control average 0.75 ± 0.02, E-row deprived average 0.74 ± 0.02, P = 0.74).
We next made experiments designed to define more precisely the time at which the mechanism operating during early postnatal development ceases to define activity-dependent changes in the cortical map. The effect of 5 days of DE pairing beginning at p9 on preferential spread of activity around stimulated D-row whiskers as measured by the C-row to E-row response ratio is indistinguishable from the effect of DE pairing from p7 to p17 (Fig. 7C). However, 5 days of DE pairing beginning at p15 resulted in temporal response profiles in C-row and E-row surround segment ROIs that were similar to control responses (Fig. 7A,B) and consequently a CE ratio time course similar to that calculated for control responses (Fig. 7C).
|
These data indicate that the mechanism that results in the observed asymmetry in the spread of subthreshold activity after DE pairing has a critical period that ends around the end of the second postnatal week. This is consistent with previous findings regarding the critical period for the effect of whisker deprivation on the appropriate development of sensory receptive fields in rat barrel cortex (Stern et al. 2001
). Alterations in Map Contours Are Transient
Monocular deprivation during the critical period for development of visual cortex results in long-term changes in network connectivity in many species including the rat (Hubel and Wiesel 1970; LeVay et al. 1980; Fagiolini et al. 1994; Pizzorusso et al. 2002), with restoration of normal function only occurring with reverse deprivation (opening the deprived eye and depriving the formerly open eye). We therefore made experiments to test whether the use-dependent alterations we observed here in somatosensory cortex were also maintained long term. Animals were DE paired from p7 to p17 with the whiskers subsequently allowed to regrow for 10 or 28 days. Nontrimmed littermates served as control animals. The mystacial whiskers regrew rapidly after the trimming had ceased, reaching approximately 50–60% of the length of the untrimmed whiskers after 10 days and approximately 80–90% untrimmed length after 28 days. After 10 days of whisker regrowth, the time course of average responses in the C-row and E-row ROIs were similar in magnitude for both control and previously DE-paired animals (Fig. 8A,B). There were no obvious differences in CE ratio temporal profiles (Fig. 8C), and CE ratios calculated from Av20to50 images were not significantly different (average CE ratio for control responses 1.03 ± 0.06, 6 D-row whisker responses from 2 control animals, average for previously DE-paired animals 0.98 ± 0.03, 14 D-row whisker responses from 5 DE-paired animals, P = 0.47). Similar results were also observed after 28 days of whisker regrowth (Fig. 8C). This result shows that the activity-dependent alterations in cortical connectivity described here are transient in nature being accommodated within 10 days. They also demonstrate a difference in the influence of afferent activity on the development of cortical connections in the visual and somatosensory cortices.
|
Speed of Activity Spread Continues to Increase after the End of the Critical Period
In addition, we observed an age-dependent effect on the time course of the spread of excitation from stimulated D-row whiskers into the surrounding rows (Fig. 9). The rate of the initial rise of the response in both C-row and E-row surround ROIs was faster in animals aged between p32 and p35 than in animals aged between p18 and p23. However, the initial rate of rise recorded from animals aged p45 to p51 was not any faster than that recorded from animals aged between p32 and p35. This result indicates that the cortical alterations that result in more rapid spread of the postsynaptic signal between adjacent cortical areas is not fully mature at the end of the critical period reported in the current study, but it continues to develop for at least a further 10 days.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsFundingReferences |
|---|
We show that pairing 2 rows of whiskers during the second and third postnatal weeks acts primarily to reduce the spread of excitation between used and deprived cortical areas. This use-dependent effect on a whiskers' cortical representation is only temporary, in contrast to use-dependent influences on the development of visual cortex where such effects are permanent (Hubel and Wiesel 1970
; LeVay et al. 1980; Fagiolini et al. 1994; Pizzorusso et al. 2002). Responses appear indistinguishable from control responses after 10 days of regrowth of trimmed whiskers and thus restoration of more normal afferent activity reaching the cortex. This effect has a critical period ending around the end of the second postnatal week, with the identical cortical deprivation protocol having no noticeable effect after p14. Laminar Locus of the Map Changes
The results suggest that the changes in vsd responses we have observed are due to changes in connectivity within the horizontal layer 2/3 network. Although we do not exclude the possibility that there are contributions from changes at other stations in the afferent pathway, such as changes in the thalamocortical projection or changes in the layer 4 to layer 2/3 projection, we regard this as unlikely. Most axons from layer 4 neurons project vertically into layer 2/3, sending relatively few collateral branches into neighboring SuW columns (Feldmeyer et al. 1999, 2002; Lübke et al. 2000, 2003; Bender et al. 2003; Bureau et al. 2004). Thus, direct activation of layer 2/3 neurons in surrounding columns is not likely to account for much of the observed vsd signal. Furthermore, when functional connectivity was assessed using laser-scanning photostimulation mapping, it was found that layer 2/3 neurons located above a barrel receive almost all their input from cells located in the underlying barrel with weak input arising from surround barrels (Shepherd et al. 2003; Bureau et al. 2004). Similar experiments assessing the spread of excitation following extracellular stimulation in layer 4 with vsd imaging in acute brain slices showed that excitation spreads into layer 2/3 in a columnar fashion (Petersen and Sakmann 2001). In acute tangential layer 4 sections, stimulation in the barrel hollow evokes vsd responses that remain restricted to the stimulated barrel and do not spread into surrounding layer 4 barrels (Petersen and Sakmann 2001), suggesting that layer 4 neurons in surrounding columns are not strongly activated by layer 4 neurons in a given principal column. Extracellular action potential (AP) recordings in vivo indicate that there is some activation of layer 4 neurons in surrounding columns at short latencies following a whisker deflection (Armstrong-James et al. 1992; Petersen and Diamond 2000). However, response magnitudes are considerably smaller than observed in the PW, and this pathway for activation of layer 2/3 neurons in surrounding columns is unlikely to account for much of the observed lateral spread of the in vivo vsd signal. Taken together, these data suggest that the major part of lateral vsd signal spread observed in vivo after a whisker deflection results from spread of excitation within the layer 2/3 network. Regarding changes in subcortical structures, experiments examining plasticity in the trigeminal ganglion and ventroposterior medial nucleus of the thalamus after whisker trimming that induces significant changes in cortical maps found that responses in the subcortical structures were unchanged (Glazewski et al. 1998; Wallace and Fox 1999). Therefore, although we do not exclude the possibility that changes in subcortical connections may account for some of the effects on the cortical responses we have observed here, changes in the layer 2/3 network appear the most likely.
Potential Mechanisms Underlying Map Changes
The second and third postnatal weeks are a period of rapid development of somatosensory barrel cortex in rats. During this period, spontaneous up- and down states increase in frequency, suggestive of progressive development and activity of the cortical network. Sensory responses acquire mature characteristics, such as shorter latencies, faster rise, and larger amplitudes, and sensory receptive fields are established (Stern et al. 2001). Morphologically, the axonal and dendritic arbors of layer 2/3 neurons expand substantially (Radnikow G, Feldmeyer D, personal communication). Behaviorally, the neonates begin whisking around p12 to p15. We show here that during this period, use-dependent mechanisms strongly effect the development of the connectivity underlying the spread of the vsd signal, acting in this case to reduce spread of excitation between neighboring used and unused cortical areas. Conceptually, this may be considered as somewhat analogous to the critical period for establishment of ocular dominance preferences in the visual cortex (Hubel and Wiesel 1970; Gordon and Stryker 1996). These sensory systems differ in that establishment of ocular dominance preference is dependent on thalamocortical afferents projecting into layer 4, whereas whisker deprivation appears here to affect primarily layer 2/3 connections. However, the result of the 2 processes is comparable in that in both cases, the presence of neighboring used and unused cortical areas results in a failure to establish connectivity or a failure of the propagation of electrical activity between used and deprived areas.
One hypothesis, referred to as the "neurotrophic hypothesis," that is advanced to explain the development of ocular dominance preferences suggests that the developing thalamocortical afferents compete for a limited supply of neurotrophic factors (Katz 1999). Active afferents or those afferents sharing correlated AP activity with neighboring synapses are proposed to enhance local production or release of the neurotrophic factor leading to stabilization of synapses and enhanced connectivity, whereas less active afferents have reduced access to the neurotrophic factor and atrophy as a consequence. It is possible to envisage a similar mechanism operating in barrel cortex. Under normal conditions of whisker usage, neighboring barrel columns receive roughly correlated activity as their whiskers' contact objects. Consequently, connections between any individual column and its neighbors are promoted approximately equally. However, upon row pairing, correlated AP activity is maintained only within the cortical columns linked to the paired rows of whiskers. In comparison, AP activity between the columns linked to spared and trimmed whiskers will become less correlated, with the consequence that "normal" network connectivity develops between used (i.e., active) columns but fails to develop between the used and the deprived columns.
Another candidate mechanism that may explain the activity-dependent changes in the spread of excitation observed here is modification to some component of the cortical inhibitory network. We find that the major change resulting from the DE-pairing paradigm is a significant reduction in the spread of the vsd response into the surrounding inactive cortical areas. This observation could also be explained by a selective increase in the inhibitory influence of the stimulated PW column on the deprived SuW columns. One recent study on the effects of deprivation on 
-aminobutyric acid synapses in rat barrel cortex found no change in synapse numbers or density in layer 2/3 (Micheva and Beaulieu 1995). The trimming paradigm used by these authors was to remove the middle 3 rows of whiskers (rows B, C, and D), leaving A- and E-rows intact. In their analysis, the authors do not differentiate between cortical columns surrounded by active SuW columns on one side and inactive ones on the other and those surrounded only by inactive columns. Changes in the number of inhibitory synapses in layer 2/3 induced by the trimming paradigm used here thus remain a possibility. In one of the few studies revealing the spatiotemporal dynamics of inhibition, Derdikman et al. show in rat barrel cortex that very strong whisker stimulation evokes net inhibition in a ring-shaped area around the stimulated whisker (Derdikman et al. 2003). This inhibitory ring covers most of the first-order SuW columns and demonstrates the potential for an inhibitory influence, at least under conditions of strong whisker stimulation, of a PW on its first-order SuWs. Consequently, a mechanism resulting in a selective increase in inhibitory potency in deprived first-order SuW columns cannot be excluded.
The results of the current work are not at first glance predicted by the conclusions of a previous study investigating changes in the short-term dynamics of intracortical synapses brought about by this deprivation paradigm (Finnerty et al. 1999). Finnerty et al. made intracellular recordings in acute slices from layer 2/3 neurons in either spared or deprived cortical columns and describe the differences observed in trains of excitatory postsynaptic potentials (EPSPs) evoked by extracellular stimulation of various intracortical pathways: the layer 2/3 pathway from the neighboring spared cortical columns, the layer 2/3 pathway from the neighboring deprived columns, and the layer 4 to layer 2/3 pathway for both spared and deprived columns. Using a model, they conclude that the changes in the trains of EPSPs are consistent with an increase in synaptic strength in the layer 2/3 to layer 2/3 pathway from spared to deprived columns and a concomitant decrease in synaptic strength in the layer 2/3 to layer 2/3 pathway from deprived to spared columns. This conclusion is also supported by minimal stimulation experiments that demonstrate that, on average, the horizontal pathway in layer 2/3 from spared-to-deprived columns has a larger initial EPSP than that from deprived-to-deprived columns. Although at first these results appear inconsistent with those of the current work, they are not necessarily incompatible. The results of the current study suggest that the deprivation paradigm results in a decrease in the net excitation spreading in layer 2/3 from the spared columns into the neighboring deprived columns. This could occur if there was a reduction in the total number of axons projecting between the columns. That said, the connections that do form between the spared and the deprived columns may be more effective than either control or deprived connections as a partial compensation. The 2 results are thus consistent, assuming that the net reduction in axonal connections was larger than the compensatory increase in individual synaptic strength.
As a final comment regarding mechanisms, it is important to note that the changes in vsd responses to whisker deflection described in this study have a clear critical period. DE pairing from p9 to p14 results in markedly asymmetric vsd responses, whereas trimming from p15 to p20 has no significant effect. However, it is clear that layer 2/3 remains plastic in adult animals, and several reports demonstrate this in rat barrel cortex (Fox 1992; Armstrong-James et al. 1994; Diamond et al. 1994; Glazewski and Fox 1996). It therefore seems most likely that these 2 forms of plasticity (i.e., the developmental, critical period-dependent form and the adult form) have as their basis separate mechanisms.
Differences to Developing Visual Cortex
One substantial difference between the developing visual and somatosensory cortices relates to the permanence of the changes in network connectivity following manipulations that alter sensory input during their respective critical periods. In visual cortex, the network connections established during this critical period persist for life. Disruption to the process by monocular deprivation (for example) leads to permanent changes in the cortical network and behaviorally to blindness (Hubel and Wiesel 1970; Katz 1999). Although the effects of monocular deprivation can be recovered or reversed under some circumstances (Movshon 1976a, 1976b; Blakemore et al. 1978; van Sluyters 1978; Kim and Bonhoeffer 1994; Liao et al. 2004; Faulkner et al. 2006), simply opening the occluded eye after closure of the critical period does not result in significant recovery of normal function (Wiesel and Hubel 1965). In contrast, in somatosensory cortex, the effects of whisker trimming on network connectivity are transient and are readily accommodated by alterations of connectivity later in development. This effect occurs within 10 days and requires only the natural regrowth of the trimmed whiskers. The fact that the somatosensory cortex continues to be malleable after the end of the critical period defined here is a feature that might be expected from a cortical area whose sensory organ can readily and rapidly fall out and regrow a multitude of times during the animals' life.
Anatomical Changes Associated with the Altered Functional Responses
Given that both axonal and dendritic arbors of layer 2/3 neurons are developing during the period of DE pairing used here, it seems possible that one or both of these arborizations may be altered. Precedents for such changes exist, for example, in visual cortex of kitten, and mouse monocular deprivation has been shown to have significant effects on axon branch pattern (Antonini and Stryker 1993, 1996; Antonini et al. 1999). A reduction in the axonal projection from the spared columns towards the deprived ones would be one possible explanation for the functional results presented here.
|
Conclusions |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsFundingReferences |
|---|
We show here that whisker trimming that leaves a used cortical area surrounded by another used area on one side and a deprived area on the other results in the development of a bias in the spread of cortical activity favoring the coactive used cortical areas. We suggest that this occurs through a mechanism that results in a weaker connectivity between the used and the unused cortical column or by a failure of normal connectivity to form between such areas. Thus, the establishment of normal connectivity between cortical areas during development may require correlated firing of APs. This effect has a critical period that ends at the end of the second postnatal week. Further, these changes in functional connectivity are not permanent, with the network retaining sufficient capacity to remodel the whisker representations in response to regrowth of the trimmed whiskers and consequent reestablishment of normal sensory input to the previously deprived area.
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsFundingReferences |
|---|
The Max Planck Society; Alexander von Humboldt Foundation.
|
Acknowledgments |
|---|
Special thanks to Thomas Hahn and Randy Bruno for frequent and fruitful discussions and to Amiram Grinvald for comments and advice during the project and critical appraisal of the manuscript. Conflict of Interest: None declared.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionsFundingReferences |
|---|
Antonini A, Fagiolini M, Stryker MP. Anatomical correlates of functional plasticity in mouse visual cortex. J Neurosci (1999) 19:4388–4406.
Antonini A, Stryker MP. Rapid remodeling of axonal arbors in the visual cortex. Science (1993) 260:1819–1821.
Antonini A, Stryker MP. Plasticity of geniculocortical afferents following brief or prolonged monocular occlusion in the cat. J Comp Neurol (1996) 369:64–82.[CrossRef][ISI][Medline]
Armstrong-James M, Diamond ME, Ebner FF. An innocuous bias in whisker use in adult rats modifies receptive fields of barrel cortex neurons. J Neurosci (1994) 14:6978–6991.[Abstract]
Armstrong-James M, Fox K, Das-Gupta A. Flow of excitation within rat barrel cortex on striking a single vibrissa. J Neurophysiol (1992) 68:1345–1358.
Bender KJ, Rangel J, Feldman DE. Development of columnar topography in the excitatory layer 4 to layer 2/3 projection in rat barrel cortex. J Neurosci (2003) 23:8759–8770.
Bernardo KL, McCasland JS, Woolsey TA, Strominger RN. Local intra- and interlaminar connections in mouse barrel cortex. J Comp Neurol (1990) 291:231–255.[CrossRef][ISI][Medline]
Blakemore C, Garey LJ, Vital-Durand F. The physiological effects of monocular deprivation and their reversal in the monkey's visual cortex. J Physiol (1978) 283:223–262.
Bureau I, Shepherd GM, Svoboda K. Precise development of functional and anatomical columns in the neocortex. Neuron (2004) 42:789–801.[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.
Diamond ME, Armstrong-James M, Ebner FF. Experience-dependent plasticity in adult rat barrel cortex. Proc Natl Acad Sci USA (1993) 90:2082–2086.
Diamond ME, Huang W, Ebner FF. Laminar comparison of somatosensory cortical plasticity. Science (1994) 265:1885–1888.
Fagiolini M, Pizzorusso T, Berardi N, Domenici L, Maffei L. Functional postnatal development of the rat primary visual cortex and the role of visual experience: dark rearing and monocular deprivation. Vision Res (1994) 34:709–720.[CrossRef][ISI][Medline]
Faulkner SD, Vorobyov V, Sengpiel F. Visual cortical recovery from reverse occlusion depends on concordant binocular experience. J Neurophysiol (2006) 95:1718–1726.
Feldman DE, Brecht M. Map plasticity in somatosensory cortex. Science (2005) 310:810–815.
Feldmeyer D, Egger V, Lübke J, Sakmann B. Reliable synaptic connections between pairs of excitatory layer 4 neurones within a single ‘barrel’ of developing rat somatosensory cortex. J Physiol (1999) 1:169–190.
Feldmeyer D, Lübke J, Silver RA, Sakmann B. Synaptic connections between layer 4 spiny neurone-layer 2/3 pyramidal cell pairs in juvenile rat barrel cortex: physiology and anatomy of interlaminar signalling within a cortical column. J Physiol (2002) 538:803–822.