By restricted use of D2 and D3 whiskers for 3–20 days at maturity (whisker pairing, WP), receptive field plasticity of adult D2 barrel cortex cells was compared in vivo for Tg8 mutant and normal (NOR) mice. Little plasticity was achieved until 20 days of WP in both mice. For Tg8, which lacks segregation of thalamocortical (TC) terminals into barrels, the first relay (TC) responses in layer IV to the principal whisker were potentiated more than in NOR mice by 20 days of WP. In parallel, secondary discharges were reduced more in Tg8 than NOR. It is suggested that both TC excitation and feed-forward inhibition in Tg8 are greater and potentiated more by WP than in NOR mice. Similar differences were reflected in supragranular (SG) cells. For Tg8 but not NOR mice, first latencies of one in five cells in layer IV to an adjacent surround whisker matched those of the principal whisker, increasing to one in three by 20 days of WP experience. Converse decreases occurred for the deprived surround whisker. Changes were similar but smaller for SG cells. Lack of TC segregation in Tg8, therefore, allows substantial overlapping TC terminals of immediate surround whiskers to activate neighbouring D2 column cells directly with potentiated relay to a whisker paired input and weakened relay to a deprived input. Although differing from NOR mice, experiential plasticity was not strongly compromised in Tg8 mice. Differences in WP plasticity from rat barrel cortex are discussed.
The monoamine oxidase-A (MAOA) knockout (Tg8) mouse exhibits a failure in segregation of thalamocortical (TC) terminals into barrel-specific clusters in layer IV of barrel cortex (Cases et al. 1996). Adult male Tg8 mice are more aggressive than wild-type mice on the same C3H background and have some problems with beam walking. However, they are physically similar and have normal brain size and gross structure (Cases et al., 1995). Due to the absence of MAOA, Tg8 mice express exceptionally high brain concentrations of serotonin transiently during the first two postnatal weeks, which prevents segregation of TC terminals and barrel formation in layer IV (Cases et al., 1996; Vitalis et al., 1998; Boylan et al., 2000). Instead, TC afferents appear evenly distributed throughout layer IV, although to similar depths as controls of the same C3H genetic background. A requirement for the disorder is a copious but transient expression of 5H-T1B receptors on TC afferents during the first postnatal week (Salichon et al., 2001).
In normal (NOR) mice, the one-to-one anatomical correspondence between each whisker and its cortical barrel (Woolsey and Van der Loos, 1970; Killackey, 1973) has provided a test-bed for unraveling TC and intracortical circuitry underpinning the somatotopic cortical map in vivo and in vitro. Quantitative analysis of spatial and temporal flow of information to and within barrel cortex has allowed considerable insight into origins of receptive fields. (Simons 1978; White, 1978, 1979; Armstrong-James and Fox, 1987; Armstrong-James and Callahan, 1991; Armstrong-James et al., 1991, 1992, 1993; Agmon and Connors, 1992; Diamond et al., 1992a,b; Welker et al., 1993; Melzer et al., 1994; Armstrong-James, 1995; Diamond, 1995; Kyriazi et al., 1996, 1998; Moore and Nelson, 1998; Rema et al., 1998; Pinto et al., 2000).
Recently, we have reported that in Tg8 mice, despite an absence of barrels and segregation of TC afferents, TC transmission in barrel cortex is more powerful than for the NOR mouse, whereas inter-columnar cortical transmission appears less efficient. (Yang et al., 2001). The present study compares experience-dependent plasticity for Tg8 and normal barrel cortex neurons and its implications for modified somatosensory circuitry. The process used involves cutting all but two adjacent whiskers in NOR mice and Tg8 adult mice on the same C3H genetic background. This innocuous technique, ‘whisker pairing’, has successfully been employed in recent years on adult rodents to investigate the roles of neurotransmitters and neurological disorders on adult cortical plasticity (Diamond et al., 1993; Armstrong-James et al., 1994; Benuskova et al., 1994; Rema et al., 1998; Sachdev et al., 1998, 2000).
Materials and Methods
Since, in rats, different barrels generate differing receptive field profiles (Armstrong-James and Fox, 1987) and the same may occur for mice, it was essential to use data only from entirely analogous cell populations. Therefore, all of the following findings concern cells exclusively in the D2 column sampled in complete penetrations through layers I–IV. All cells were tested to D2 (principal) and immediately adjacent D1 and D3 whiskers for comparative assessment of centre and surround receptive field (SRF) data.
Topographical Organization of Barrel Cortex in Tg8 mice
Since barrels are absent in Tg8 mice, histological identification of different columns is prohibited. However, in an earlier study (Yang et al., 2001), receptive fields and temporal responses for barrel cortex cells were compared in NOR and Tg8 mice to determine whether any differences might reflect known anatomical differences. It was found that radial columnar segregation and topographical organization is as good, statistically, in Tg8 mice as NOR mice. Here, we have used the identical procedure for finding the D2 ‘column’ in NOR and in Tg8 mice, which is briefly as follows. The first microelectrode penetration in any experiment was placed ~3.5 mm lateral and 1.5 mm caudal to bregma where D2 columns are usually found (Yang et al., 2001). If this revealed successive cells maximally responsive to the C2 whisker, the next penetration would be 200 μm further medial, where the expected site of a D2 column would be in a NOR mouse. If in the new track cells were responsive maximally to D1 or to C2 or to D3, then the next penetration would be moved rostral, medial or caudal, respectively, to the expected site of a D2 column. This procedure invariably was successful in Tg8 as well as NOR mice, since the D2 column was found always within 200–300 μm of an adjacent D1, D3, C2 or E2 column in both Tg8 and in NOR mice.
Successful neurophysiological investigations of receptive fields (RFs) of D2 column cells in barrel cortex were carried out on 43 fully adult male C3H mice (mean wt 27.65 ± 0.21, range 25–30 g) and 40 fully adult male Tg8 mice on a C3H background (mean wt 28.43 ± 0.47, range 24.3–32 g). Postnatal ages for each group ranged between 9 and 15 weeks at onset of first procedure. For experimental groups, all whiskers except D2 and D3 were cut on the right side of the face, while those on the left were left intact. This condition is termed ‘whisker pairing’, here defined as WP experience (Diamond et al., 1993; Armstrong-James et al., 1994). ‘Acute’ whisker trimming was carried out for a period of 3 days to be compared with ‘chronic’ whisker trimming for 20 days. NOR littermates with intact whiskers were used for comparison [n = 17 controls (NOR); n = 15 Tg8]. While the mice were gently hand-held immobile, all but D2 and D3 whiskers on the right were clipped to the level of the fur. Whisker paired mice were housed two to four littermates in a cage over 3 or 20 days of pairing, preceding the physiological recording sessions. For 3 days of WP experience, whiskers were cut once and allowed to re-grow for 72 h before experiments were performed (n = 16 NOR; n = 14 Tg8). For the 20 days WP group, whiskers were cut every other day and allowed to re-grow for the last 72 h before physiological recording (n = 10 NOR; n = 11 Tg8). The numbers of cells in each subgroup for which full data were available are given in Figure 1 (insets for each figurine). Thirty minutes prior to recording, the two intact whiskers were trimmed to a length of 3–5 mm to match the length of the previously clipped whiskers. This enabled equivalent deflections in stimulation trials for each whisker. For WP experience, clipped whiskers were defined as ‘deprived whiskers’; intact whiskers were defined as ‘maintained whiskers’.
Anaesthesia and Surgery
Mice were surgically anaesthetized by the i.p. route with urethane (10% solution in distilled water; 1.5 mg/g body wt). Anaesthesia was maintained such that animals did not whisk or show a blink reflex and pinching of the hind foot failed to evoke a withdrawal reflex. If an animal showed any one of these signs an extra injection of urethane (10% of the initial dose) was given i.p. Body temperature was maintained at 36–37°C by a thermistor-controlled heating pad (Harvard plc, Leeds, UK). Mice were placed in a headholder (Kucera, 1970) and exposed to a continuous epinasal flow of oxygen for subsequent surgery and neurophysiology. The left parietal cortex was exposed by craniotomy using a small electric drill. The dura mater was left intact and irrigated with phosphate-buffered saline continuously.
Whisker stimuli were identical for all whiskers tested. To deflect individual whiskers on the right side of the face, a piezoelectric ceramic wafer stimulator was placed just below the shaft of the whisker ~4 mm from the skin. The wafer was deflected by a computer-gated trapezoid pulse with rise and fall times each of <0.5 ms and a total onset-offset duration of 3 ms. The stimulus deflection at the probe tip was 300 μm up–down. The stimulating probe was carefully observed using a dissection microscope, to ensure it did not touch adjacent whiskers and skin.
Single unit extracellular recordings were made with tip-etched carbon-fibre microelectrodes (Armstrong-James and Millar, 1979; Armstrong-James et al., 1980). Penetrations were at ~30° from vertical, normal to the dura to allow orthogonal cell sequencing from one column. Contact was verified through a dissection microscope (Zeiss). These microelectrodes easily penetrate dura mater with negligible dimpling (Armstrong-James et al., 1992). Sub-pial penetration was registered by an appearance of neuronal noise, from which locus intracortical depths were measured.
Successful search coordinates for the D2 column (see below) ranged from 1.2 to 2.0 mm caudal to the bregma, 3.5–4.0 mm lateral to the midline and 0.15–0.4 mm from the pial surface; all cell depths were logged. A calibrated three-dimensional stepping-motor microdrive (Narishige, Tokyo, Japan) with an operational accuracy of ±10 μm was used for electrode displacement.
Cell Sampling and Analysis
Spike signals were enhanced by band-pass filtering at 0.8–8 kHz using Neurolog spike-processing modules (Digitimer, Welwyn Garden City, UK). Single cells were discriminated by multiple component discrimination of amplitude and duration of negative and positive spike components using an in-house waveform-window multiple level discriminator generating TTL logic, time-stamped output. Sampling resolution for each data channel was set at a 1/50 μs interval. Spike amplitudes of less than five times baseline noise were rejected for sampling; only cells with initial negative-going action potentials were used.
If the first two encountered cells responded maximally to the D2 whisker, then all cells down to a depth of at least 425 μm (deep layer IV) were investigated for responses to all three whiskers. To each whisker (D1, D2 or D3), post-stimulus time histograms (PSTH) and modal latency histograms were generated on-line from the discriminator output, using a CED-1401 processor and Spike 2 software (Cambridge Electronic Design, Cambridge, UK) with in-house programs. For each block of stimulus trials, each whisker was stimulated 50 times at 1 s intervals. At least 40 s was allowed to elapse between trials. Each cell was subjected to latency and response magnitude (RM) analysis from PSTH data from the D2 (principal) and D1 and D3, (surround) whiskers. The RM was defined by the total number of spikes generated 300 ms post-stimulus to 50 whisker deflections minus equivalent period spontaneous activity (Armstrong-James and Fox, 1987). Spontaneous activity was derived from mean values immediately prior to stimulation sequences. Latency histograms were assembled from each first spike generated post-stimulus in trials. Analysis was restricted to responses from cells located at depths of 0–425 μm spanning the first four cortical layers. Cells were allocated for the separate analyses within the SG layer (0–270 μm depths) or the granular layer IV (>270 μm to <425 μm). All data were displayed and stored via a Pentium PC for further data abstraction and subsequent statistical analysis off-line.
Procedures were identical to an earlier study on Tg8 and NOR mice (Yang et al., 2001). Briefly, at the end of recording sessions small lesions were made at sites in the cortex from which recordings were made, using 1.5 μA of negative current through the electrode. Subsequently, the brain was removed and fixed in 4% paraformaldehyde solution for subsequent sectioning. Sections were stained in cresyl violet or with the cytochrome oxidase technique for visualization of lesioned sites and barrels. In common with our earlier studies on adult mouse barrel cortex (Welker et al., 1993, 1996), we found that the barrel junction between the SG layer and layer IV extended over 250–300 μm below the pial surface (Yang et al., 2001). The layer IV/Va junction was at 400–450 μm. Accordingly, cell populations were divided into two groups, SG layer cells being nominated as those found within 270 μm of the pial surface and layer IV cells as those located between 270 and 425 μm of the pial surface. Although Tg8 mice exhibit no distinct barrels, the laminar delineation of layers I–IV appeared normal in confirmation of a previous study (Cases et al., 1996). Consequently, for Tg8 animals similar depths were used for classifications of layer boundaries.
Complete response and latency data on 333 cells in NOR mice and 397 cells in the Tg8 mice were gathered for comparisons of responses to 50 stimuli applied to the center-receptive ‘principal’ D2 whisker and to each of the adjacent SRF whiskers in the same row; D1 and D3. Cells were rejected for analysis if any data were absent for latency or response to each SRF and principal whisker. For NOR mice, 178 and 155 cells were classified as being located in layer I–III (SG layer) and layer IV, respectively, of the D2 column (see Materials and Methods). For Tg8 mutant mice, 230 and 167 D2 column cells were classified as being within the SG layers and layer IV, respectively. Breakdown for numbers of cells in each sub-group are to be found in the insets to Figure 1.
Experience-dependent Changes in RMs
Figure 1 shows changes in total RM with 3–20 days of WP experience for D1, D2 and D3 whisker receptive field components of D2 column cells for NOR and Tg8 mice, respectively. For the SG cell population in Tg8 mice, no changes in RM occur for any whisker with 3 or 20 days WP experience. For SG cell populations in NOR mice, differences in RM to the D2 principal whisker between controls and either 3 or 20 days of WP experience were not significant [P > 0.15, Mann–Whitney U-test (MWU)]. Similarly, for the D3 whisker in NOR mice, differences between control and whisker paired values for SG cells were not significant for 3 or 20 days of WP experience (P > 0.10, MWU). However, for the D1 deprived surround whisker in NOR mice there is a highly significant fall in response to ~50% of control values after 20 days of deprivation for SG cells (P < 0.003, MWU), although changes were not significant after 3 days of WP experience (P > 0.2, MWU).
In common with the SG layers in NOR mice, RMs of layer IV cells to the paired D2 and D3 whiskers were unchanged by WP experience, (D2 whisker, P > 0.15, MWU; D3 whisker, P > 0.05, MWU). RMs to the D1 (deprived) whisker, however, showed a precipitous decline in layer IV (P < 0.0001, MWU) in common with SG layers at 20 days of WP experience. By contrast, in Tg8 mice RMs for layer IV cells showed no significant changes in total response with 3 or 20 days of WP experience for any whisker (P > 0.15 for D2 whiskers, P > 0.06 for D1 or D3 whiskers).
Changes in Spontaneous Activity
Since spontaneous firing rates in part reflect mean levels of depolarization of cells and thereby influence RMs to afferent inputs, they were recorded for every cell studied. Mean spontaneous firing rates for the stock of cells analysed in Figure 1 are shown in Figure 2 separately for SG and Layer IV populations.
Mean spontaneous firing rates were considerably higher in NOR than Tg8 mice for both SG (NOR 1.61 ± 0.16, Tg8 0.62 ± 0.06); and layer IV (NOR 1.81 ± 0.15, Tg8 0.31 ± 0.05) cell populations for the control animals. With increasing WP experience (0–3–20 days) spontaneous firing rates declined substantially for SG layer cells of both NOR and Tg8 cell populations, this effect being more significant for NOR cells.
In NOR mice, spontaneous firing rates of layer IV cells declined significantly to similar levels at 3 and 20 days. In layer IV of Tg8 mice, however, alterations in spontaneous activity levels with 3 or 20 days of WP experience were not consistent or significant.
Changes in Temporal Coding of Responses
In previous studies on rats we have found that plasticity for the principal (D2) whisker response can be masked by opposing changes in short latency or long latency components (epochs) of post-stimulus histograms in cortex (Diamond et al., 1993; Armstrong-James et al., 1994). Those findings also were used to differentiate direct TC from intracortically relayed discharges. In Figure 3A,B, RMs to D2 are reconfigured by condensing PSTH data into sequential ‘PSTH-epochs’ to discriminate short from longer latency discharges with WP experience.
Figure 3A compares, for SG layers, changes in temporal coding of the principal whisker D2 responses with increasing WP experience through 3 to 20 days, for the cells in the SG layers of NOR (left column) and Tg8 mice (right column). The greatest changes occur over the first 20 ms of the response. In the control condition, earliest discharges ( ms post-stimulus) to D2 are more prevalent in Tg8 than NOR mice. For NOR, the firing frequency of earliest discharges potentiates only at 20 days of WP. For Tg8 mice, frequencies of earliest discharges also potentiate with increase in duration by 20 days of WP experience. Intermediate latency discharges (10–20 ms) decline in tandem. In essence, it is these opposing changes which result in null changes in overall RMs to D2 for SG cells in Tg8 and NOR mice (Fig. 1).
Figure 3B shows changes in temporal coding of D2 whisker responses for layer IV cells of NOR and Tg8 mice. Cells were encountered in the same penetrations as SG cells in Figure 3A. Again, the greatest changes occur over the first 20 ms of the response and more early discharges to D2 occur in Tg8 mice at any stage than for NOR mice. In common with SG cells, firing rates in this earliest epoch (<10 ms) progressively potentiate with duration of WP experience, whereas in the intermediate latency epoch (10–20 ms) firing rate reduces in parallel. Again, these opposing changes in transmission underlie null changes recorded for total RMs to the principal whisker, D2, for layer IV cells in Tg8 and NOR mice (Fig. 1).
Figure 4A allows detailed and statistical comparison of changes in mean response within the earliest PSTH epoch (<10 ms) for SG layers and layer IV for NOR and Tg8 animals. Potentiation only reaches significant levels with prolonged WP experience (20 days) for both NOR and Tg8 animals when differences from controls are highly significant (P < 0.005, MWU). For both Tg8 and NOR mice, potentiation of response within this earliest epoch is greater for layer IV than the SG layers.
Figure 4B shows the evolution of intermediate PSTH epochs (10–20 ms post-stimulus) with WP experience. Note that changes are the reciprocal of those for earliest discharges (A,B). Again, with short WP experience (3 days) changes are not significant, whereas with prolonged WP experience (20 days) reductions of response in this epoch are highly significant in both NOR and Tg8 mice (layer IV NOR P < 0.0005, MWU; other incidences P < 0.005, MWU). Reductions in the intermediate response with WP experience for layer IV cells is greater than for SG layer cells for both types of mice.
Latencies to the Principal Whisker
Latencies of the first spike of each response were collected for all cells for all whisker deflections in the present study.
Figure 5 documents the changes in values of median latencies of responses to the principal whisker input, D2. Latency changes by 3 days of WP were not significant for either SG or layer IV cells in any mice (not shown). However, for both NOR and Tg8 mice, mean latencies were decreased at 20 days of WP experience. Changes were significant for all layers studied in both Tg8 and NOR mice. Latency shortening was least for SG layers in Tg8 mice and greatest for SG layers of NOR mice.
The way in which median latencies to the principal D2 whisker shorten with experience is revealed in Figure 6, where the influence of 20 days of WP experience on median latency distributions is shown by comparison with control data. Two trends are apparent: (i) for controls and for both periods of whisker pairing, short latencies (6–8 ms) were considerably more abundant in Tg8 mice than NOR mice; this was so for all layers; and (ii) the incidence of short latencies increased progressively with duration of WP experience in all layers, pari passu, with the elimination of longer latency responses; this process occurred for both Tg8 and NOR mice.
Surround Whisker Latencies
For SRF whiskers, evoked spikes in controls were generated at much more distributed and over longer latencies than were responses to the principal whisker. PSTH-epoch analysis, therefore, produced inadequate data for useful statistical analysis. However, first spike latency analysis produced useful findings for responses to surround whiskers.
Figure 7 compares average median latencies to D1 and D3 surround whiskers for controls and 20 days of WP experience. In the absence of WP experience in control animals, the differences in latency to the D1 and D3 surround whiskers were not significant for layer IV or SG cells. This was also the case for NOR or Tg8 animals (P > 0.10 in all cases, WMPSR).
However, in common with principal whisker responses, latencies to the deprived D1 and maintained D3 whiskers changed with 20 days of WP experience in both NOR and Tg8 mice for D2 column cells. Changes at 3 days of WP experience were not significant (not shown). However, for Tg8 animals, whisker pairing for 20 days caused the latency to D3, the paired whisker, to shorten very significantly relative to the deprived D1 whisker for both SG and layer IV cells (Fig. 7, right column). For NOR mice, changes for D3 were much less (Fig. 7, NOR, left column) and differences between the deprived D1 and maintained D3 surround whiskers less significant. It is clear that this form of plasticity was greater for Tg8 than NOR animals in both layers tested.
Short Latency SRF Responses
Responses in barrel cortex at <10 ms latency are suspect direct TC inputs (Armstrong-James and Fox, 1987; Armstrong-James and Callahan, 1991; Armstrong-James et al., 1991, 1992) operating principally through Ampa-kainate receptors (Armstrong-James et al., 1993). A separate analysis of these short latency responses for SRF inputs, therefore, may give insights into changes in more direct TC inputs as against intracortical or other circuitous relays to the D2 column.
For NOR mice, the incidence of short latency responses to adjacent D1 or D3 SRF inputs in the control condition was rare, being <3.5% for layer IV cells and <5% for SG cells (Fig. 8, upper). After 20 days of depriving the D1 whisker of NOR mice, every SG cell or layer IV cell tested failed to respond to this whisker at short latency (<10 ms). On the other hand, for the paired D3 whisker a modest increase to just >10% of cells fired at latencies of <10 ms in SG cells, but to only 4.2% for layer IV cells.
By contrast with NOR mice, short latency SRFs were common in Tg8 mice under control conditions (Fig. 8, lower filled histograms). Following 20 days of WP experience in Tg8 mice, the proportion of layer IV cells responding at short latency to the paired D3 whisker was increased to 31.3% from 18.4% for controls (Fig. 8, lower left). Conversely, for the deprived D1 whisker the incidence of short latency responses fell from 26.5% (controls) to 15.7% in layer IV. Similar, but more dramatic changes occurred for SG cells in Tg8 mice (Fig. 8, lower right). For the D3 paired whisker, short latency responses effectively doubled in incidence from 15.7% of cells in controls to 30.6% with whisker pairing. In stark contrast, for the deprived D1 whisker only 2.1% of cells responded at short latencies compared with 18.9% for the D1 whisker in controls.
Are Fast SRFs Big SRFs?
Although, in Tg8 mice, many cells fired at short latencies to SRF inputs and incidences increased to the paired whisker, we wanted to know if responses to these inputs were large or small and, hence, whether the balance of RMs was influenced by experience.
Figure 9 compares, for Tg8 mice, the distributions of RMs for the cells described for Figure 8 firing at short median latencies (<10 ms post-stimulus). For layer IV, the relative incidence of larger responses to the D3 whisker is increased more than for small responses following pairing for 20 days. By contrast, for the deprived D1 whisker smaller responses predominated for short latency responses.
For SG cells, pairing the D3 whisker for 20 days caused a novel appearance of some large responses, but in addition produced a very substantial increase in very small short latency responses (Fig. 9, middle right histogram). Responses are too few for the deprived D1 whisker for SG cells to allow comment (Fig. 9, lower right). Finally, for NOR mice changes in fast SRFs also were too small to be analysed.
In Tg8 mice the lack of clustering of TC terminals into whisker-specific cortical barrels might be expected to underlie some differences in receptive field plasticity of NOR and Tg8 mice. However, as reported earlier, receptive fields of both Tg8 and NOR mice are dominated by strong principal whisker responses and low amplitude surrounds (Yang et al., 2001). Here, we have found that whisker pairing plasticity induces differential changes in short latency and secondary relay for centre (principal) and surround receptive fields. Changes in these components and their possible origins are discussed separately below.
In rats, short latency responses (<10 ms post-stimulus) in barrels to principal whiskers are driven through AMPA-kainate receptors (Armstrong-James et al., 1993) and these responses are commensurate with direct TC relay (Armstrong-James and Fox, 1987; Armstrong-James et al., 1992). In the present study, potentiation of principal responses was limited to analogous short latency TC discharges. In both types of mice, their potentiation was double the control level at 20 days of WP for layer IV, but always greater in Tg8 mice where control responses also were larger. The latter may be promoted by an exuberant growth of TC fibres in layer IV (Vitalis et al., 1998). However, from a visual examination of dextran-labeled TC fibres, their collateralization in layer IV appears no more dense in Tg8 mice (Cases et al., 1996).
Potentiation of the earliest principal whisker responses could result from (i) more early spikes being generated by a fixed proportion of cells, or (ii) more cells producing short latency responses. From latency distribution analysis, WP experience caused a higher proportion of cells to fire at short latencies to the principal whisker in both types of mouse. This suggests either that some novel excitatory TC synapses were generated on normally unresponsive cells, or that existing ‘silent’ or inadequate TC synapses (Feldman et al., 1999) have been upregulated. Nothing is known of relative synaptology for Tg8 and NOR mice.
Potentiation of early responses of SG cells occurred in parallel with layer IV changes for both Tg8 and NOR mice. This is to be expected, since to principal inputs responses of SG cells are driven powerfully by intra-column relay from layer IV (Armstrong-James et al., 1992), although some TC fibres do invade layer III in both rats and mice (Killackey, 1973; Keller et al., 1985; Lu and Lin, 1993). Potentiation for SG cells of Tg8 mice was less than for NOR mice and less than occurred for layer IV cells in the same penetrations. Similarly, latency changes to principal whiskers with pairing were least for SG cells in Tg8 mice and greatest for SG cells of NOR mice (Fig. 5). Together, these findings suggest deficient plasticity of intra-columnar excitatory transmission of centre receptive field data from layer IV to upper layers in Tg8 mice, whereas plasticity of TC relay is more potent in Tg8 than NOR mice.
Feed-forward Excitation and Inhibition
Virtually all sensory responses following initial TC relay in barrel cortex are likely to be compounded of competing excitatory and inhibitory post-synaptic potentials (Armstrong-James and Fox, 1987; Armstrong-James et al., 1992; Agmon and Connors, 1992). In the present study, we found that secondary (10–20 ms latency) responses were depressed in both NOR and mutant mice by 20 days of WP experience. As a general observation, the depression of secondary responses mirrored the potentiation of earliest responses in all layers in both types of mice.
Secondary excitatory responses most likely arise by intracortical relay through recurrent excitation of primary discharge of neighbouring cells (Douglas et al., 1995). The timing of the peak depression of responses at 10–20 ms post-stimulus is co-temporaneous with that of feed forward in-field afferent inhibition (Laskin and Spencer, 1979; Gardner and Costanzo, 1980), which peaks at 5–10 ms after peak TC excitation in layer IV of S1 rodent cortex (Armstrong-James and George, 1988; Agmon et al., 1996, Kyriazi et al., 1998, Moore and Nelson, 1998). Together, the above findings satisfactorily account for depression of secondary discharge to principal whisker stimulation being mirrored by potentiation of excitatory TC responses. Correspondingly, the most profound reduction of responses at 10–20 ms post-stimulus occurred for layer IV cells in Tg8 mice, for which potentiation of primary discharge was maximal. In short, the simplest explanation is that in-field afferent inhibition is dictated by primary excitatory drive, in particular in layer IV and, hence, potentiation of excitation and inhibition go hand in hand.
The net effect of the above experience-dependent changes in excitation and inhibition was to focus responses to shorter latencies in both the SG layers and layer IV. The focusing process is more profound in Tg8 mice than NOR mice in all layers studied and perhaps arises from the consistently faster relay of TC inputs for these mice. Whether TC synapses are more profuse on layer IV cortex neurons of Tg8 mice than NOR mice remains to be investigated. However, our findings give no reasons to believe that the physiological processes underlying experience-dependent plasticity of principal whisker responses differ fundamentally in Tg8 and NOR mice.
In the rat, under similar conditions of testing and anaesthesia to those used here, SRFs in layers I–IV have been shown to depend for their expression on intracortical column-to-column relay of principal whisker discharges by a number of criteria (Armstrong-James and Callahan, 1991; Armstrong-James et al., 1991; Fox, 1994, Wallace and Fox, 1999; Fox et al., 2001). Furthermore, plasticity of SRFs in layers I–IV of barrel cortex has been shown to be dependent on local cortical circuitry. WP potentiation is prevented by focal suppression of intracortical transmission in rat barrel cortex through NMDA receptors (Rema et al., 1998) and whisker deprivation plasticity by local barrel cortex muscimol block (Wallace et al., 2001), leaving VPM thalamic relay potency unaffected. It is unknown if mice follow similar rules for SRF generation as rats. However, this is likely to be the case for our NOR mice, since without whisker pairing the differences in mean latencies between SRF and principal whisker responses were some 15–20 ms, in common with rat barrel cortex.
WP experience caused a decrease in latency to paired whisker stimulation of both layer IV and SG layer cells in Tg8 mice (Fig. 7). Conversely, latencies to deprived whisker responses increased, suggesting either excitatory synaptic weakening or net potentiation of background inhibition. Much smaller changes in the same directions were seen for NOR mice. In these respects, plasticity was greater for Tg8 mice.
The distribution of latencies for SRFs in control Tg8 animals revealed an unusual minority of very short latency (<10 ms) responses, which were absent in NOR mice (Fig. 8). In Tg8 mice, the incidence of these fast responses increased markedly for the paired SRF whisker and decreased for the deprived SRF whisker. However, in NOR mice, these fast responses only made a minor appearance in layer IV with whisker pairing, being otherwise absent. We suggest that very short latency SRF responses in layer IV of Tg8 mice arise from some inappropriate TC terminals from D1 and D3 barreloids targeting the D2 column in layer IV. This would satisfactorily explain why latencies of these minority fast SRF responses were no different from those to the principal whisker. It also accounts for why the paired surround D3 whisker response was potentiated in a similar way to principal whisker responses. None the less, these mis-targeted SRF collaterals still generated low individual powers compared to longer latency responses to the same whisker. Their mean RM also was small relative to the principal whisker, ensuring preservation of discrete somatotopic targeting in the face of TC cluster overlap. In rats subjected to 30 days of WP experience, some short latency paired SRF whisker responses commensurate with monosynaptic TC relay also appear for barrel cells (Armstrong-James et al., 1994), although these are less common than found here for Tg8 mice.
Deprived SRF inputs have less opportunity for generating correlated firing with intact inputs due to a lack of use (Benuskova et al., 1994). In NOR mice, responses to the deprived (cut) surround whisker after 20 days of cutting were strongly reduced for both the SG and layer IV cells, relative to same whisker responses in control mice. These findings were entirely similar to those found in our previous studies on normal rat barrel cortex cells (Armstrong-James et al., 1994). However, for Tg8 mice, deprivation was less effective. No changes in overall mean RMs for deprived surround whiskers were seen, either in SG layers or layer IV, although for both layers the occurrence of short latency responses (<10 ms) was reduced.
Serotonin and Plasticity
Although Tg8 mice express exceptionally high (2- to 9-fold normal) brain concentrations of serotonin during the first two postnatal weeks (Cases et al., 1995, 1996), levels are profoundly reduced at maturity, in part due to the presence of the MAOB enzyme (Cases et al., 1995; Ikemoto et al., 1997).
Although functional presynaptic 5-HT1 B receptors exist in neonates, they are almost completely redundant in adult mouse barrel cells (Laurent et al., 2002). A significant effect upon TC transmission by serotonin in NOR and Tg8 adult mouse barrel cortex is, therefore, unlikely. Iontophoresed serotonin does depresses S1 spontaneous and evoked activity in layers other than IV (Waterhouse et al., 1986; Bassant et al., 1990), but direct effect on layer IV cells seems unlikely, since in layer IV in mouse barrel cortex no post-synaptic action is detectable (Laurent et al., 2002). For cells in SG layers, it remains uncertain whether serotonin concentrations in adult Tg8 brains would significantly affect spontaneous or evoked activity.
Comparisons with Rat WP Plasticity
In both the Tg8 and NOR mice studied here, the absolute magnitudes of responses to principal and paired whiskers failed to augment through 3 and 20 days of whisker pairing. However, in rats, whisker pairing typically results in some 30 and near 100% potentiation of paired principal and surround whisker responses, respectively, in the targeted barrel-column (Armstrong-James et al., 1994). Furthermore, WP plasticity is detectable much earlier, within 1 day of WP experience in rat barrel-column SG cells (Diamond et al., 1994) and within 3 days, peaking at 7 days, for barrel cells (Armstrong-James et al., 1994).
In rat barrel cortex early (3 days) WP plasticity to principal whisker stimulation only occurs for responses relayed secondarily (>10 ms latency) and not for primary TC relay (Armstrong-James et al., 1994). This contrasts strongly with both Tg8 and NOR mice, where significant potentiation only occurred for TC-relayed responses and only with periods of >3 days of whisker pairing.
In rats, extending WP experience from 3 to >7 days causes a reversal from potentiation to reduction of secondary intracortical discharges to the principal whisker (Armstrong-James et al., 1994). In our Tg8 and NOR mice the analogous responses were not potentiated at all. Instead, they were reduced at 3 and 20 days of WP experience. This suggests that plasticity of afferent inhibition is greater in both types of mice than in rats or, alternatively, that the efficacy of excitatory neuronal plasticity is lower. Low plasticity of excitatory intracortical relay in our mice is also suggested by their deficient SRFs relative to rats. D1 and D3 SRFs for D2 barrels in control normal rats are about twice the magnitude of those seen in the control NOR and Tg8 mice used here (Armstrong-James et al., 1994; Rema et al., 1998).
Levels of spontaneous neuronal activity also impact on RF size and plasticity. Experiments (Ho and Destexhe, 2000) have shown that increased spontaneous activity in cortical neurons allows response to otherwise sub-threshold inputs. Mean spontaneous activity was low in Tg8 mice compared with NOR mice and in both types of mice spontaneous activity was reduced progressively by whisker pairing experience, profoundly so in Tg8 mice (Fig. 2). In rats, spontaneous activity levels were higher at all broadly equivalent stages of control and WP experience than for either mouse described here (Armstrong-James et al., 1994). Low spontaneous activity would compromise NMDAr activation on which intracortical sensory relay depends (Armstrong-James et al., 1993) and for which ongoing post-synaptic depolarization is required (Bindman et al., 1988; Agmon et al., 1996).
We thank Dr Egbert Welker and Dr Patricia Gaspar for helpful discussions on these experiments. This work was supported by a European Union Grant (Biomed 2) BMH4 CT-97-2412.