Spontaneous bursting and rhythmic activity in the cuneate nucleus of anaesthetized rats

Neuroscience 141 (2006) 487–500

SPONTANEOUS BURSTING AND RHYTHMIC ACTIVITY IN THE CUNEATE NUCLEUS OF ANAESTHETIZED RATS E. SÁNCHEZ, A. REBOREDA, M. ROMERO AND J. A. LAMAS*

1968; Brown et al., 1974; Amassian and Giblin, 1974; Canedo et al., 1998). Moreover, spontaneous rhythmicity is characteristic of cuneate neurones in anesthetized cats. Although single-spiking activity can be observed, both evoked and spontaneous firing have a tendency to appear in the form of high frequency bursts that contain several spikes (Galindo et al., 1968; Amassian and Giblin, 1974; for a review see Canedo, 1997). Importantly, both cuneolemniscal cells and cells presumed to be cuneate interneurones display spontaneous rhythmic firing in the cat, at least in the slow (⬍1 Hz), delta (1– 4 Hz) and gamma (30 – 80 Hz) like frequencies (Mariño et al., 1996, 2000; Canedo et al., 1998). To date, the spontaneous activity of cuneate neurones has not been deeply studied in anesthetized rats, and it has been suggested that these neurones may be similar to the gracile neurones (Panetsos et al., 1998; Nuñez et al., 2000). In rats, neurones of the gracile/cuneate nuclei have been classified as high-frequency “interneurones” (HF or type I; 14 Hz rhythm) and low-frequency projection cells (LF or type II; non-rhythmic ⬍10 spikes/s), mainly on the basis of their single-spiking spontaneous activity (Panetsos et al., 1997, 1998; Nuñez et al., 2000). In rat slice preparations, cuneate neurones have also been classified in two types with similar, but not identical, characteristics to those reported “in vivo” (Nuñez and Buño, 1999). To our knowledge, spontaneous bursting activity has not been described in neurones of the rat dorsal column nuclei (DCN) in vivo. However, intrinsically driven clustering activity (spontaneously repeated groups of two to five action potentials at about 1 Hz) was recently demonstrated in cultured rat DCN neurones with a rhythmic low threshold membrane potential oscillation of about 10 Hz being a common characteristic (Reboreda et al., 2003). Furthermore, cells showing spike frequency adaptation and spontaneously rhythmic single action potentials were also detected in culture. The complex behavior observed in anesthetized cat CN neurones and cultured rat DCN neurones, but also the intricate composition of the rat DCN in terms of cellular and neurotransmitter types (Popratiloff et al., 1996, 1997), suggest that the activity of the rat cuneate neurones in vivo might be more diverse than previously reported. In agreement with this hypothesis, we show here that cuneate neurones in anesthetized rats display a complex spontaneous activity. Spontaneous rhythmic and random activity was readily observed in the form of bursts and/or single action potentials. Silent cells, although less common, were also detected when electrical shocks were applied to the medial lemniscus or upon hair brushing to the forelimb.

Physiology Section, Department of Functional Biology, Faculty of Biology, University of Vigo, Lagoas-Marcosende, 36310 Vigo, Spain

Abstract—Spontaneous and rhythmic neuronal activity in dorsal column nuclei has long been identified in anesthetized cats. Here, we have studied the spontaneous behavior of cuneate cells in anesthetized rats through extracellular recording, showing that most cuneate neurones recorded (155 of 185) fired spontaneously. Overall, 74% of these spontaneously firing neurones were single-spiking and 26% were bursting. Cells were considered “bursting” when more than 50% of the spontaneous spikes belonged to bursts. Nevertheless, occasional bursts were seen in 33% of spontaneous cuneate cells which were classified as single-spiking. Rhythmic firing was observed in about 14% of both spontaneously bursting and single-spiking cells, and these cells were located close to the obex (ⴞ0.5 mm). Although the spike-frequency was mostly in the range 0 –15 spikes/s, spontaneous rhythmic activity was circumscribed mainly to the alpha/betalike range, both in single-spiking (26.1ⴞ3.6 Hz, nⴝ16) and bursting cells (19.5ⴞ4.1 Hz, nⴝ6). Lemniscal stimulation often activated several antidromic units with the same latency. About 65% of cuneolemniscal cells were spontaneously active and of these, 83% were single-spiking and 11% rhythmic (all single-spiking). In cells that were not antidromically activated from the medial lemniscus, short latency orthodromic responses consistent with excitation by recurrent lemniscal collaterals were often observed following lemniscal activation. Interestingly, only cells completely unresponsive to lemniscal stimulation showed rhythmic bursting. Most spontaneous cells responded with a burst to natural receptive field stimulation, while rhythmic cells became temporally arrhythmic. These results demonstrate, for the first time, that rat cuneate neurones can fire bursts spontaneously. Besides, this bursting activity can be rhythmic. These two properties, and the fact that groups of cuneolemniscal cells share the same conduction velocity, probably imply the reinforcement of temporal and spatial summation at their targets when they are synchronously recruited by the stimulation of overlapping receptive fields. © 2006 IBRO. Published by Elsevier Ltd. All rights reserved. Key words: excitability, oscillation, dorsal column nuclei, cuneolemniscal, electrophysiology, anesthetized rat.

The firing pattern of neurones in the cuneate nucleus (CN) has been studied extensively in anesthetized cats in response to peripheral afferents and sensorimotor cortex stimulation (e.g. Towe and Jabbur, 1961; Galindo et al., *Corresponding author. Tel: ⫹34-986-812563; fax: ⫹34-986-812556. E-mail address: [email protected] (J. A. Lamas). Abbreviations: CN, cuneate nucleus; DCN, dorsal column nuclei; EGTA, ethylene glycol-bis-(b-aminoethylether)-N,N,N,N-tetraacetic acid.

0306-4522/06$30.00⫹0.00 © 2006 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2006.03.050

487

488

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

Thus, some of the rhythmic events reported here may reflect intrinsic membrane properties and indicate an active role of CN in sensory transmission. Some of these data were previously reported in abstract form (Sánchez et al., 2004).

EXPERIMENTAL PROCEDURES Anesthesia and surgery Adult Sprague–Dawley rats (320.9⫾11.8 g) were pre-anesthetized with ether (Panreac Química S.A., Barcelona, Spain). After receiving an i.p. dose of ketamine⫹xylazine (40 mg/kg⫹4.6 mg/ kg; Sigma-Aldrich Chemie, Steinheim, Germany) or pentobarbital (50 mg/kg, Sigma-Aldrich Chemie), an endotracheal cannula was inserted in order to artificially ventilate the animal (49 strokes/min; 1.5–2 ml/stroke) and to measure and control the end-tidal CO2. A cannula was inserted into the jugular vein and a mixture of anesthetics (ketamine 25 mg/h⫹xylazine 3 mg/h or pentobarbital 30 mg/h) and a paralyzing agent (gallamine triethiodide, 1 mg/h; Sigma-Aldrich Chemie) was continuously administered using an analgesic administration pump (Sorenson Medical, West Jordan, UT, USA). No significant differences were found in the behavior of cells recorded under different anesthetics (see Discussion). Anesthesia was controlled by monitoring the cardiac rhythm and pupil diameter. The animal was fixed to a stereotaxic device with the head slightly flexed to improve access to the DCN, and local anesthetic (Xylocaine: lidocaine 2% gel) was topically administered to pressure points. In order to gain visual access to the dorsal surface of the CN, the cisterna magna was carefully opened and the lower part of the occipital bone and the dorsal arch of the atlas were removed. Bilateral pneumothorax was routinely carried out to avoid respiratory movements and the exposed nervous tissue was covered with a thin layer of 4% agar prepared in physiological saline. At the end of the experiment, the animal was killed by administering an overdose of anesthetic. All procedures received prior approval from the Spanish Research Council and the University of Vigo Scientific Committees, and they fulfill all Spanish and European directives (R.D. 223/1988 and 86/609/EEC). Every effort was made to minimize the number of animals used and their suffering.

Recording and stimulation To record extracellular activity from single neurones in the CN, sharp glass microelectrodes (50 –100 M⍀) filled with K-acetate or Na-citrate (2 M) were used. Alternatively, patch-clamp like micropipettes were used (5–10 M⍀), filled with an intracellular solution containing (in mM): potassium acetate 90, KCl 20, MgCl2 3, CaCl2 1, EGTA 3 and HEPES 40, pH adjusted to 7.2 with NaOH 20 mM. No differences were observed using recording electrodes of different types. Most electrode tracks were carried in four antero-posterior transversal planes, where the density of cuneothalamic neurones is the highest (Barbaresi et al., 1986; Kemplay and Webster, 1989). These planes were located at the following positions with respect to the obex: 0.5 mm rostral (⫹0.5), at the obex (0.0), 0.5 mm caudal (⫺0.5) and 1 mm caudal (⫺1). Some electrode tracks were also carried out in a plane 1 mm rostral (⫹1) for comparison. Most penetrations were made between 1 and 1.5 mm lateral to the midline, and at a depth of up to 1 mm. Electrode penetration, was carefully controlled using a stereoscopic microscope and an electrically driven micromanipulator (Conmot 250, Cibertec, Madrid, Spain). Note that we considered the obex as the point where the gracile fasciculi begin to diverge and the caudal end of the area postrema appears at the dorsal midline (Barbaresi et al., 1986). An Axoclamp 2B (Axon Instruments Inc., Foster City, CA, USA) or a Dagan 2400 (Dagan Corporation, Minneapolis, MN,

USA) amplifier was used to amplify the signals recorded, the recordings being stored in a digital tape for further analysis (sampling rate 48 kHz; DTR-1205, Bio-Logic Science Instruments, Claix, France). All signals were band-pass filtered between 0.1–3 kHz or low-pass filtered (DC-5 kHz) and additional digital filtering was used where necessary. The signals were visualized online through an oscilloscope and a computer monitor. Extracellular action potentials were mainly negative when the electrode was relatively distant from recorded cells, but often they became positive (positive–negative) as the electrode approached the cell. This latter shape was considered to be a quasi-intracellular recording. Similar observations have been reported previously for cat gracile and cuneate neurones (Towe and Jabbur, 1961). A bipolar concentric stimulating electrode was placed at the medial lemniscus (A4.5, L2, H3; Paxinos and Watson, 1998) and electrical stimulation through this electrode was used as hunting stimulus in all experiments (one to 10 stimuli/s). Electrical shocks of 0.1 ms duration and 0.1–1.5 mA amplitude at variable frequencies were delivered through photoelectric stimulus-isolation units (Grass S-88 and PSIU6, Grass Instrument Division Astro-Med Inc., West Warwick, RI, USA). Cells responding to medial lemniscus stimulation were considered antidromic when the latency was constant, followed electrical stimulation at more than 100 Hz and the collision test was positive (see Lamas et al., 1994). These cells were classified as cuneolemniscal and they were roughly considered as cuneothalamic. Independently of the cell response to the lemniscal stimulation, care was taken in order to record all spontaneous cells appearing during electrode tracks, this allowed the construction of Table 1. To assess the correct positioning of recording electrodes at the CN, a handheld brush and/or a hand-operated air-jet stimulus through a Pasteur pipette, was directed toward the ipsilateral forelimb. Movement of the forelimb around joints was also carried out.

Data analysis Data were analyzed and plotted off-line using Spike2 and Origin5 software (Cambridge Electronic Design Limited, Cambridge, UK; Origin Laboratory Corporation, Northampton, MA, USA). In order to study the spontaneous discharge, histograms for inter-spike intervals (interval histograms) were plotted. When a short latency peak appeared in the histogram, this was used to determine the maximum inter-spike interval to be considered as belonging to a burst (always less than 10 ms). Bursts that fulfilled this criterion were searched and analyzed. Overall, if more than 50% of the spikes belonged to bursts the cell was considered as bursting. Periodicity was assessed and the rhythm frequency obtained by analyzing autocorrelograms. Bursts or single-spikes were considered rhythmic only when the autocorrelogram resulted in three or more clear peaks. Note that even in a rhythmic single-spiking cell, the action potential firing frequency (spikes/s) may not be identical Table 1. All cells recorded (n⫽355) Firing activity

Spontaneous Rhythmic Nonrhythmic Subtotal Silent Unknown

Firing type

Subtotal

Bursting

Single-spiking

6 (4%) 35 (23%) 41 (26%)

16 (10%) 98 (63%) 114 (74%)

155 22 (14%) 133 (86%) 30 170

“Unknown” cells were antidromically identified but not further recorded; they were used in the latency histogram of Fig. 4C.

E. Sánchez et al. / Neuroscience 141 (2006) 487–500 to the frequency of the rhythm (Hz), mainly because firing may occasionally fail. Averages are given as the mean⫾S.E.M.

RESULTS The extracellular data reported here were obtained from a total of 355 cuneate neurones, a pool of neurones that was divided into several subgroups attending to different characteristics. The latency to medial lemniscus stimulation was the only information extracted from one group of 170 cells, as the records were not long enough for any further analysis to be performed. In contrast, when recordings were long enough (ⱖ1 min; n⫽185), the cells could be classified as spontaneous or silent neurones (see Table 1). Among these 185 neurones, spontaneous activity was observed in 155 cells whereas 30 cells were considered silent (only revealed by lemniscal stimulation). Spontaneous cells were classified in function of their spontaneous activity as bursting (n⫽41) or single-spiking (n⫽114). The cells within each of these groups could be further broken down into rhythmic or non-rhythmic neurones (Table 1). Bursting cells fired repetitively in groups of two (doublets) or more action potentials, with inter-spike intervals inside each burst of less than 10 ms. Singlespiking cells consistently fired single action potentials at different frequencies. Cell firing was considered rhythmic when the “autocorrelogram” showed at least three clear peaks of activity. Note that some cells showed occasional bursts mixed with single-spikes (see Amassian and Giblin, 1974), but they did not fulfill the criteria described in Experimental Procedures and as such, were considered singlespiking cells (n⫽51). Hence the percentage of spontaneous cells showing bursts was 59% (92/155): 33% showed occasional bursts and 26% fired bursts consistently. A group of neurones (n⫽106 of 172) were driven by electrical stimulation of the medial lemniscus and were classified as lemniscal (71; cells responding antidromically) or orthodromic (35; cells responding orthodromically). A further 66 of the 142 spontaneous neurones tested did not respond to this stimulus at all (non-responding cells). The spontaneous firing of the cells in each of these groups was classified as above (Table 2). Distribution of recorded cells Neurones were recorded from 1 mm rostral to 1 mm caudal to the obex in five standard transversal planes: ⫹1, ⫹0.5, 0.0 (obex), ⫺0.5 and ⫺1 mm with respect to the obex (see Fig. 1 and Tables 3 and 4). When a deviation from these standard rostro-caudal recording planes had to be made to avoid blood vessels, the neurones recorded were assigned to the closest standard plane. In each plane, cells were recorded mainly from 1 to 1.5 mm lateral to the midline at depths of up to 1 mm. Spontaneously active cells. Single-spiking neurones were found to be the most abundant in all five planes and they were homogeneously distributed in the rostro-caudal direction (83, 71, 76, 70 and 71% of the total spontaneous cells recorded in planes: ⫹1, ⫹0.5, 0.0, ⫺0.5 and ⫺1, respectively). Hence, the percentage of bursting cells

489

Table 2. Cells classified in function of their response to the lemniscal stimulation and type of spontaneous firing Response

Firing activity

Firing type Bursting

Antidromic (lemniscal) (n⫽71) Orthodromic (n⫽35)

Not affected (n⫽66)

Spontaneous (n⫽46) Rhythmic Nonrhythmic Silent (n⫽25) Spontaneous (n⫽30) Rhythmic Nonrhythmic Silent (n⫽5) Spontaneous (n⫽66) Rhythmic Nonrhythmic

Single-spiking

0 8

5 33

0 7

3 20

6 16

5 39

ranged between 17 and 30% of the total spontaneously active cells recorded in these standard planes (Table 3). A 2⫻5 contingency table, extracting “single-spiking” and “bursting” columns from Table 3, was constructed and analyzed. The resulting chi-square value was 0.75, indicating that there is not a statistically significant relationship between both variables (position and type of activity) at the 0.05 level (chi-square critical value⫽9.49). It is worth noting that although the number of cells recorded at ⫹1 was low, five out of the six were single-spiking. It may be relevant that rhythmic cells were only found in the planes ⫹0.5, 0.0 and ⫺0.5 (20, 11 and 15% of the total spontaneous cells, respectively), indicating that spontaneously rhythmic cells tend to lie around the obex. Spontaneous cells that were antidromically activated also showed a tendency to concentrate around the obex, although the distribution was not statistically significant at the 0.05 level (0, 26, 36, 40 and 18% of the total spontaneous cells, respectively). Lemniscal cells. The group of lemniscal cells was distributed between ⫹0.5 and ⫺1 mm from the obex (Table 4), antidromic cells were not found at ⫹1 level (see Table 3). On closer examination, we found that the proportion of silent lemniscal cells clearly increased in the more caudal planes with respect to the total number of antidromic cells recorded in each plane (18, 26, 43 and 73%, respectively). A 2⫻4 contingency table (rows: silent and spontaneous cells; columns: planes ⫹0.5, 0, ⫺0.5 and ⫺1) produced a chi-square value of 9.6, indicating a significant relationship between the proportion of silent cells and their position at Pⱕ0.025 (critical value⫽9.35). In contrast, antidromic single-spiking cells were mostly located around the obex (64, 62, 50 and 18%, respectively; although this tendency was not statistically significant at the 0.05 level) and antidromic bursting cells did not express a clear localization (18, 12, 7, and 9%, respectively; the small number of bursting cells precluded chi-square testing). Much like the spontaneous rhythmic cells, all rhythmic lemniscal cells were found around the obex, in planes ⫹0.5, 0.0 and ⫺0.5 (9, 9 and 7%, respectively). All percentages and tests considered the total lemniscal cells recorded in each plane.

490

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

Fig. 1. Distribution of sampled cuneate neurones. Location of the neurones recorded with respect to the obex (0.0), “⫹” means rostral and “⫺” means caudal. The squares represent silent cells, circles represent single-spiking cells and triangles are bursting cells. Filled symbols represent cells antidromically activated by medial lemniscus stimulation. Note that when two or more cells shared exactly the same position they were slightly displaced in order to make them visible.

These data highlight three clear points: 1. when considering spontaneous cells, firing in bursts is not related to the location of the cell along the rostrocaudal axis, 2. rhythmic cells are located close to the obex, and 3. when considering lemniscal cells, the proportion of silent antidromic neurones clearly increases in the caudal direction, while spontaneous antidromic cells seem to express a preference for locations around the obex. Spontaneous bursting cells So far, from the data we have analyzed some 26% (41/ 155) of the spontaneous cells recorded were bursting neu-

rones (Fig. 2A1 and 2B1; see also Table 1). The inter-spike interval histogram of bursting cells had a very distinctive early peak at less than 10 ms (shorter intra-burst intervals; Fig. 2A2 and 2B2). A second group of intervals (longer inter-burst intervals) was also observed which resulted in a very clear and narrow (Fig. 2A2) or a very broad and inconspicuous (Fig. 2B2) peak. The presence of two clear and narrow peaks was characteristic of rhythmic bursting cells (4%; six of 155) while the interval histograms showing a clear early peak but a broad or non-existent second peak corresponded to bursting, non-rhythmic cells (23%; n⫽35/ 155). Consistent doublets, such as those reported classi-

E. Sánchez et al. / Neuroscience 141 (2006) 487–500 Table 3. Rostro-caudal distribution of the number of spontaneous cells with respect to the obex (Plane 0) Plane

Single-spiking

Bursting

Rhythmic

Antidromic

Total

⫹1 ⫹0.5 0 ⫺0.5 ⫺1

5 (83%) 25 (71%) 53 (76%) 14 (70%) 12 (71%)

1 (17%) 10 (29%) 17 (24%) 6 (30%) 5 (29%)

0 7 (20%) 8 (11%) 3 (15%) 0

0 9 (26%) 25 (36%) 8 (40%) 3 (18%)

6 35 70 20 17

Percentages are given in relation to the total spontaneous cells recorded in each plane.

cally in anesthetized cat cuneate neurones, were found in 14 of the 41 bursting cells (see Fig. 2A1). The autocorrelograms of rhythmic bursting cells showed at least three clear peaks. Indeed, in the example shown in Fig. 2A, seven clear peaks were observed with a 0.5 s scale, giving a rhythmic bursting frequency of about 12 Hz (Fig. 2A3). In contrast, autocorrelograms of nonrhythmic bursting cells showed a unique and conspicuous peak at less than 10 ms resembling intra-burst intervals (as shown in Fig. 2B3). Only 15% (six of 41) of the bursting cells showed a clear rhythm. It is relevant to mention here that while spontaneous doublets and/or longer bursts are characteristic of the cat cuneate neurones (Amassian and Giblin, 1974; Canedo et al., 1998), to our knowledge this is the first report of such a behavior in rat cuneate neurones. Spontaneous single-spiking cells Single-spiking cells constituted 74% of the spontaneous cells recorded (114/155; Table 1), 16 of which were rhythmic (10%; 16/155; Fig. 3A1). Rhythmic cells showed a clear peak in their interval histograms and some of them presented a smaller second peak due to occasional firing failure, which resulted in intervals with durations twice that of the main interval (Fig. 3A2). Their autocorrelograms were characterized by the absence of a peak below 10 ms and that seen in Fig. 3A3 shows a rhythm of about 18 Hz. No clear peaks were observed in the interval histograms of spontaneous non-rhythmic single-spiking cells (63%; 98/ 155; Fig. 3B1 and Fig. 3B2) or in their autocorrelograms (Fig. 3B3). Furthermore, it should be noted that the percentage of single-spiking cells being rhythmic (14%; 16/ 114) was similar to that of bursting cells. Antidromic lemniscal activation The effect of electrical stimulation of the contralateral medial lemniscus was tested in a total of 342 cells. Of these, 241 were activated antidromically, 35 orthodromically and 66 remained unaffected. The action potentials evoked antidromically had an invariable latency, collided with spontaneous orthodromic activity (Fig. 4A) and followed high frequency electrical stimulation (100 Hz or higher; Fig. 4B). The mean antidromic latency from 241 neurones was 1.7⫾0.04 ms (Fig. 4C), yielding a mean conduction velocity of 7 m/s over an average distance of 12 mm from the stimulating site to the CN; distance was estimated follow-

491

ing the course of the cuneolemniscal tract in the stereotaxic atlas (Paxinos and Watson, 1998). In 71 of the 241 cuneolemniscal cells, the recordings were maintained for sufficient time to investigate their firing pattern. Of these, 46 (65%) were spontaneously active and 25 were silent (35%; Table 2). Most spontaneous cuneolemniscal cells were single-spiking and non-rhythmic (72%; 33/46), while 11% (5/46) showed single-spiking rhythmicity. Although eight of the spontaneous lemniscal cells were bursting (17%; 8/46) none of them were rhythmic, which could indicate that rhythmic bursting cells may be interneurones. Overall, only five of 71 (7%) lemniscal cells were rhythmic and all of these were single-spiking. It is noteworthy that 32% (46 of 142) of the tested spontaneous cuneate neurones were antidromically activated from the medial lemniscus. The probability of any cuneothalamic axon being spared by the stimulus was low. Indeed, not only was the stimulating intensity well above threshold for most of the antidromically evoked responses but also, the axons in the medial lemniscus are tightly packed. In two of these 46 spontaneous antidromic cells, lemniscal stimulation induced firing arrest following antidromic activation (see next section and Fig. 5D). Lemniscal stimulation often activated groups of antidromic units that showed the same latency (ie: the same conduction velocity). Upper trace in Fig. 4D shows the progressive activation of three neurones when the stimulating intensity was increased from subthreshold for the cell “a,” to a suprathreshold level for all the three cells (a⫹b⫹c). The fact that this activity was from three different neurones was demonstrated using the appropriate collision tests with spontaneous action potentials (Fig. 4D, middle and lower traces). Furthermore, similar phenomena have been reported in cat cuneolemniscal cells (Canedo, 1997) and in several motor tracts (e.g. pyramidal: Canedo and Towe, 1985; rubrospinal: Canedo and Lamas, 1989; reticulospinal: Canedo and Lamas, 1993). Orthodromic lemniscal activation and unresponsive cells Non-antidromic effects were observed in 35 cells following lemniscal stimulation (Fig. 5 and Table 2) and these effects could be divided in three main categories: activation, inhibition and slight changes in spontaneous firing frequency. Orthodromic activation consisted in the firing of one, two or a burst of action potentials with a variable but relatively consistent latency in peristimulus histograms (Fig. 5A–C). Cells firing only one or two orthodromic spikes tended to Table 4. Rostro-caudal distribution of the number of cuneolemniscal cells with respect to the obex (Plane 0) Plane

Silent

Single-spiking

Bursting

Rhythmic

Total

⫹0.5 0 ⫺0.5 ⫺1

2 (18%) 9 (26%) 6 (43%) 8 (73%)

7 (64%) 21 (62%) 7 (50%) 2 (18%)

2 (18%) 4 (12%) 1 (7%) 1 (9%)

1 (9%) 3 (9%) 1 (7%) 0

11 34 14 11

Percentages are given in relation to the total lemniscal cells recorded in each plane.

492

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

Fig. 2. Bursting neurones. (A) Example of a cuneate spontaneous bursting cell firing doublets (1). The interval histogram shows two clear peaks (2) and the autocorrelogram indicates the firing of rhythmic bursts at about 12 Hz (3). (B) Another spontaneously bursting cell (1) showing only one clear peak in the interval histogram (2). The autocorrelogram indicates the firing of non-rhythmic bursts (3). Bin size: 1 ms.

have relatively conserved latencies (Fig. 5A, B), while the shortest latency in those neurones firing a burst was more variable (Fig. 5C). In two spontaneous cells, lemniscal stimulation caused an arrest in their firing for more than 200 ms, as was seen in other two cuneolemniscal cells (Fig. 5D). The firing properties of a group of 66 spontaneous cells, which were not affected by lemniscal stimulation, were also studied (see Table 2). It is worth noting that neither lemniscal nor orthodromic cells fired rhythmic bursts, unlike six cells that did not respond. Although the percentage of spontaneous cells that were rhythmic was also higher in the last group (11, 10 and 17% respectively), this difference was not statistically significant at the 0.05 level (chi-square test and Fisher exact test). Bursting activity was more often observed in spontaneous non-responding (33%) than in spontaneous lemniscal (17%) or orthodromic cells (23%), but again these differences were not statistically different at the 0.05 level. Firing frequency Although spontaneous cuneate cells showed a wide range of firing frequencies (0.02– 65.1; mean: 9.7⫾1 spikes/s),

they were mostly below 15 spikes/s (123 of 155, 79%; Fig. 6A and 6B). It is also worth noting that using a 1 Hz bin width, the highest value appears in the bin of zero to one spike/s (32 of 155, 21%), indicating that cuneate neurones fire within a low frequency range. Bursting cells have two other frequencies which deserved to be investigated. The intraburst frequencies ranged from 125 to 574.7 spikes/s with a mean of 313.4⫾18.5 spikes/s (n⫽41; Fig. 6C). On the other hand, the mean interburst frequency was 3.7⫾0.9 bursts/s (range 0.03– 22.7), with most lying in the range of zero to six bursts/s (34 of 41, 83%; Fig. 6D). Again the highest bar is that of the bin zero to one (40%). The distribution of spike frequencies for antidromically activated cuneate neurones (Fig. 6E) was no different from that of the whole population (mean: 7.1⫾1.2 spikes/s; range: 0.1–31.7). Most of them fell in the range 0 –15 spikes/s (40 of 46, 87%) and a good amount (12 of 46, 26%) in the bin of zero to one spike/s. These results indicate that spontaneous cuneate cells tend to fire, spontaneously, between 0 and 15 spikesbursts/s independently of being cuneolemniscal or noncuneolemniscal neurones.

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

493

Fig. 3. Single-spiking neurones. (A) Example of a spontaneous single-spiking cuneate cell (1). The interval histogram shows two peaks, the second one due to occasional firing failure (2), while the autocorrelogram indicates a rhythmic firing at about 18 Hz. (B) Spontaneous single-spiking neurone (1) with no conspicuous peaks either in the interval histogram (2) or in the autocorrelogram (3), indicating a non-rhythmic firing. Bin size: 1 ms.

Rhythm frequency

Sensorial modality

On the whole, 14% (22/155) of the spontaneous cuneate cells recorded fired rhythmically (see Table 1). According to the autocorrelograms, rhythmic events were detected in both single-spiking and bursting cell groups, with most cells (16 of 22) firing in the range of alpha/beta-like frequencies (10 –30 Hz). A clear rhythm was observed in 14% of the single-spiking cells (16/114; Fig. 6F; dashed bars), with a mean frequency of 26.1⫾3.6 Hz (range 11.3– 64.4 Hz). Bursting cells also fired rhythmic bursts (15%; six of 41), at a mean frequency of 19.4⫾4 Hz (range 11.2– 34.2 Hz; Fig. 6F; filled bars). In those cells from which longer-lasting basal recordings were obtained (more than 3 min), the possibility of a slow-like rhythm (⬍1 Hz) was investigated using 10 s autocorrelograms. None of the nine bursting (0 rhythmic) and 32 single-spiking cells (six rhythmic) analyzed showed a clear rhythm at these frequencies. Finally, it is worth noting that while the rhythmic events in bursting cells involve a burst of several spikes as opposed to a single-spike, the frequency range was basically the same, alpha/beta-like, in both cell types.

It was not the aim of this study to investigate the characteristics of the receptive fields in depth; however, the modality of a group of recorded neurones was manually determined. Cells were classified as responding to brief air-jets, brief taps or to movement around joints. All stimuli were applied to the ipsilateral foreleg, neck or shoulder. A total of 49 single-spiking neurones (of 75 tested) were activated by peripheral stimulation (18 were lemniscal and seven were rhythmic). Overall 14 responded to air-jets (five lemniscal, one rhythmic), 17 to taps (seven lemniscal, three rhythmic) and 18 to movement (six lemniscal, three rhythmic). Thirty single-spiking cells (18 lemniscal, three rhythmic) responded to adequate stimulus with a short burst (Fig. 7A), eight cells (three lemniscal, two rhythmic) responded with a short burst followed by an inhibition period (Fig. 7B), five cells showed inhibition only (Fig. 7C; one lemniscal, one rhythmic) and six cells (three lemniscal, one rhythmic) showed a subtle change in frequency. Also 19 of 28 non-rhythmic bursting cells tested responded to peripheral stimulation: five to air-jets (one lem-

494

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

Fig. 4. Lemniscal antidromic activation. (A) A cuneate neurone showing constant latency to repetitive electrical stimulation (three superimposed traces). Note the collision of the antidromically evoked action potential when the stimulus is preceded by a spontaneously activated orthodromic action potential. (B) The cell in A following 1:1 high frequency (100 Hz) lemniscal stimulation. (C) Histogram showing the antidromic latency for 241 lemniscal neurones. Bin size: 0.2 ms. (D) Increasing the intensity of lemniscal stimulation shows the superposition of three different units (a, b and c) responding antidromically with the same latency (above). Collision of the antidromic action potential from cell c with a spontaneous action potential is shown in the middle panel. Collisions obtained when stimulating at subthreshold intensities for cell c allow the antidromic activity of cells a and b to be distinguished (lower panel). In this and the subsequent figures, electrical stimulus artifacts are marked with asterisks.

niscal), 10 to taps (five lemniscal) and four to movement (two lemniscal). Seventeen showed a short burst (eight lemniscal) and two showed a burst followed by inhibition. In all four bursting rhythmic cells (two tap, two movement) receptive-field stimulation induced the disorganization of the rhythm. Fig. 8A shows the response of the bursting-rhythmic cell described in Fig. 2A to the movement of the ipsilateral foreleg, note that the spontaneous rhythmic activity was resumed 0.5–1 s after the stimulation. Comparison of the autocorrelogram shown in Fig. 8B (obtained during stimulation) with that in Fig. 2A3 (obtained from spontaneous activity) indicates that sensorial stimulation induced the disorganization of the rhythm in this representative cell.

In summary, the percentage of single-spiking cells responding to different stimuli (air-jets, taps and movement) was similar, and most of them (78%) responded to their adequate stimuli with a burst of action potentials. Bursting cells preferred brief-taps and also most of them (83%) fired a burst in response to receptive field stimulation. Interestingly, all bursting rhythmic cells responded to the stimulation losing their rhythms and only one of the 11 spontaneously rhythmic cells tested responded to air-jets.

DISCUSSION This study has revealed several important findings concerning the spontaneous and rhythmic activities of cuneate

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

495

Fig. 5. Lemniscal evoked orthodromic activity. (A) Electrical stimulation of medial lemniscus induced a response with variable latency in a cuneate cell (right, five traces superimposed). Although variable, a peristimulus raster (left and above) and histogram (left and below) indicate a relatively consistent latency at about 6 ms. (B) Another cell showing a double response to lemniscal stimulation (right, five traces superimposed). The peristimulus raster and histograms (left) showed that the first action potential has a less variable latency that the second one. (C) Cuneate cell responding with a burst to lemniscal stimulation (right). The raster plot and histogram on the left indicate the highly variable onset latency. (D) In a few spontaneous cells, lemniscal stimulation caused firing arrest for long periods as shown by the long lasting recording (right) and by the peristimulus raster and histogram (left). Note that in this figure the bar (in histograms) and data points (in rasters) closest to zero represent stimulus artifacts, except in part D, where they represent the antidromic activation of the cell. Bin size: 1 ms.

496

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

Fig. 6. Spike frequency and rhythm frequency. (A) Spike frequency histogram for 114 single-spiking cells. Bin size: 1 spike/s. (B) Spike frequency histogram for 41 bursting cells. Bin size: 1 spike/s. (C) Frequency of action potentials inside the bursts for 41 bursting cells. Bin size: 50 spikes/s. (D) Frequency of bursts for bursting cells. Bin size: one burst/s. (E) Spike frequency for 46 antidromically activated cuneate neurones. Bin size: 1 spike/s. (F) Rhythm frequency histogram for rhythmic bursting (black) and single-spiking (dashed) cells. Bin size: 10 Hz.

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

497

Fig. 7. Effects of receptive-field stimulation in single-spiking lemniscal cells. (A) Spontaneous activity (1) of a single-spiking, low-frequency cell responding with bursts (2) to air-jets applied repetitively to the dorsal forepaw (2). (B) Spontaneous activity (1) of a cell responding with a burst followed by a period of inhibition (2) to the stimulation with brief taps applied to the fourth finger. (C) Spontaneous activity (1) of a neurone inhibited (2) by the movement of the ipsilateral foreleg around the wrist. Because the stimuli were applied manually, arrowheads only approximately indicate the start of the stimulus. All three neurones were antidromically activated by lemniscal stimulation.

neurones in anesthetized rats. About 26% of spontaneous cuneate neurones can be considered to be bursting cells, but an even higher percentage (about 60%) showed bursts. Although spike firing frequency was more often found in the 0 –15 spikes/s range, spontaneous rhythmic

activity was found between 10 and 60 Hz in both bursting and single-spiking neurones (mostly in the alpha/beta-like range). About 65% of cuneolemniscal neurones fired spontaneously and, when antidromically activated, axons of cells located physically close conducted at the same ve-

Fig. 8. Effects of receptive-field stimulation in bursting-rhythmic cells. (A) Response of a rhythmic bursting cell to the movement of the ipsilateral forelimb. Asterisk approximately marks the start of the stimulus. (B) The autocorrelogram of the stimulus-evoked activity shows the absence of rhythmicity when compared with the autocorrelogram of spontaneous activity for the same cell (see Fig. 2A3 for comparison). Bin size: 1 ms.

498

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

locity. Although some lemniscal cells were bursting, rhythmic activity was not observed in any of these. In contrast, six non-cuneolemniscal neurones (putative interneurones) were bursting and rhythmic. Natural receptive field stimulation evoked a burst of action potentials in most spontaneous cells; however, in rhythmic cells the rhythmic pattern was momentarily disrupted under stimulation. Spontaneous activity, bursting behavior and rhythmicity are common characteristics of cuneate neurones in anesthetized cats (for a review see Canedo, 1997). Our study demonstrates that these features are also applicable to cuneate cells in anesthetized rats, although differences between both species may be also relevant. Lemniscal stimulation The mean antidromic latency of cuneolemniscal cells was 1.7 ms and although the position of the stimulating electrode may be slightly different, this value is similar to the 1.5 ms reported by Davidson and Smith (1972) under similar conditions. In our study, lemniscal stimulation often activated several antidromic responses with the same latency but belonging to different cuneolemniscal cells. This observation reinforces the anatomical data that suggest the presence of tightly packed groups of cells (Basbaum and Hand, 1973; Barbaresi et al., 1986; Kemplay and Webster, 1989), perhaps similar to the “clusters” in the main CN of cats (see Maslany et al., 1991). Furthermore, it means that cells which are close together, and that probably share a similar receptive field, have an almost identical conduction velocity. This characteristic might be important to reinforce the transmission of information to the thalamus as it favors a good spatial summation if the neurones are recruited to fire synchronously (see Canedo, 1997). Along with their bursting activity, this could be the basis of the positive cross-correlation between pairs of cuneate and thalamic cells observed when their receptive fields are stimulated (Alloway et al., 1994). Under such conditions, bursting cuneate cells with the same conduction velocity can be recruited together, maximizing spatial and temporal summation at their targets. However, it was noteworthy that when two clearly different spontaneous cuneate cells were simultaneously recorded through the same electrode, no positive cross-correlation was found between them (14 pairs tested). Indeed, similar results were obtained in rat DCN neurones by Nuñez et al. (2000), suggesting that sensory stimulation is necessary to phaselock the firing of physically close cuneate neurones. Lemniscal stimulation induced orthodromic responses in 35 of 342 cells tested. It is known that lemniscal stimulation can induce three types of activity in cat cuneate neurones (Canedo et al., 2000): antidromic responses, orthodromic activity through recurrent collaterals of cuneolemniscal cells and orthodromic activity through a thalamo-cortical loop. In cat cuneolemniscal neurones, antidromically activated spikes are often followed by a second orthodromic spike due to the synaptic impact of recurrent collaterals from other cuneothalamic cells (Aguilar et al., 2002). However, it has also been proposed that recurrent collaterals only establish direct synaptic contact

with interneurones in the cat and rat CN (Andersen et al., 1964; Davidson and Smith, 1972). In this study, putative recurrent activity was not observed in any cuneolemniscal cell concomitant with the antidromic response. Besides, it has been recently shown in cats that the recurrent effect is mostly seen only when the stimulating electrode is located in the minimum threshold site for antidromic activation of the recorded cell, and/or when a GABA antagonist is applied to the CN (Aguilar et al., 2002). Neither of these maneuvers was attempted here. It is also possible that basal GABA inhibition (lateral recurrent inhibition) is stronger in the rat than in the cat CN, a question that should be addressed in the future. However, we did find significant orthodromic activation in non-lemniscal cells and while in some, the latency of these responses was consistent with a recurrent effect, in others the long latency could be better explained by a cortical loop. Firing frequency Although some silent cells might have been passed unnoticed in this study, it is generally accepted that most cuneate neurones are spontaneously active in anesthetized animals (Amassian and Giblin, 1974; Brown et al., 1974; Pubols et al., 1989; Alloway et al., 1994; Canedo et al., 1998; this study). To our knowledge, no bursting behavior has previously been described in rat cuneate cells; however, most spontaneous cuneolemniscal feline neurones fire in bursts (about 68%; Canedo et al., 1998), whereas presumptive interneurones mostly show single-spike firing (about 80%; Canedo et al., 2000). Similarly, 74% of spontaneous cuneate neurones fired bursts in anesthetized raccoons (Pubols et al., 1989). Here is shown that about 26% of spontaneous cells could be considered as bursting cells, additionally, 33% showed occasional bursts so that spontaneous cells showing bursts numbered 59%. These data demonstrated that spontaneous bursting activity in the DCN is not exclusive of carnivores. A striking difference was that only about 17% of antidromically identified spontaneous lemniscal cells could be considered as bursting neurones in our study. However, if we add the cells that showed occasional bursts but that were not considered bursting (total⫽51, antidromic⫽23), the percentage of the spontaneous lemniscal cells showing bursts reaches 67%, close to the proportions reported in carnivores. In our study, only those cells in which at least 50% of action potentials were registered inside bursts and having an intraburst frequency of more than 100 Hz were considered bursting neurones. We are unaware of the criteria established in other studies. This apparently arbitrary criterion was used to avoid considering “bursting” those cells which fired mainly single-spikes and only occasionally a burst, we rationalized that an occasional burst could be due to uncontrolled afferent activation (most cells responded with bursts to receptive field stimulation), while a consistent spontaneous bursting might have a strong intrinsic component. Regardless, the percentage of spikes inside bursts of those cells considered single-spiking was always well below 50% (mean⫽25%). Furthermore, we cannot rule out differences between species regarding intrinsic membrane properties,

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

synaptic connectivity or both. Indeed, intraburst frequencies reached up to 575 Hz in agreement with very short (⬍4 ms) burst interspike intervals found in cat and raccoon cuneate neurones (e.g. Galindo et al., 1968; Pubols et al., 1989). Spontaneous single-spiking activity has been reported in neurones of the rat gracile nucleus (Panetsos et al., 1997) and DCN (Panetsos et al., 1998; Nuñez et al., 2000). In those studies, most DCN neurones were silent or fired at frequencies below 10 spikes/s. Our results confirm that most single-spiking cells fire at low frequencies, in agreement also with a mean of 4.9 spikes/s (range 0.5–30 spikes/s) reported for rat cuneate neurones by Alloway et al. (1994). Importantly, our study also shows for the first time that the burst frequency (interburst) remains below six bursts/s in most bursting cells (mean: 3.8 bursts/s). Previously, the low frequency and the silent cells were considered as cuneolemniscal cells, while presumed interneurones fired at more than 10 spikes/s (but see Nuñez and Buño, 1999). Here, we show that the distribution of spike frequencies for cuneolemniscal cells is very similar to that of the whole population. Moreover, we observed a similar spike frequency distribution for single-spiking cells (8.7⫾0.9 spikes/s) and bursting cells (12.4⫾2.6 spikes/s; see Fig. 5), as reported for cuneate cells in the raccoon (Pubols et al., 1989). These results suggest that the bursting/non-bursting behavior, but not the overall spike frequency, may be the functional attribute that best characterizes cuneate neurones. High frequency bursts should allow temporal summation at the level of cuneothalamic synapses to occur and hence, to increase the power of single cuneothalamic neurones. Rhythmicity Both lemniscal and non-lemniscal cuneate neurones show resting rhythms at slow, delta and gamma frequencies in anesthetized cats (Mariño et al., 1996; Canedo et al., 1998). As reported by Amassian and Giblin (1974), who found that only 17% of spontaneous cat cuneate neurones were rhythmic, only a small percentage of rat spontaneous cuneate cells are rhythmic (14%); however, the range of rhythmic frequency in rats was basically limited to the alpha/beta-like range. Nevertheless, it is important to bear in mind that the spike frequency was principally found in the 0 –15 spikes/s range. In support of these results, betalike rhythmicity has been reported in multi-unit recordings from rat DCN neurones (Panetsos et al., 1998). It remains to be determined why spontaneous rhythms below 10 Hz are absent in anesthetized rats. However, our results from spontaneous single-spiking cells are in accordance with the rhythmic (23 Hz) EPSPs recorded in rat ventro-posterolateral thalamic neurones, previously demonstrated to be originated by the regular firing of prethalamic DCN neurones (Pinault and Deschenes, 1992). One of the major afferent input to the CN comes from corticofugal neurones (see Martínez et al., 1995) and it has been suggested that the cuneate slow rhythm may be imposed from the cortex in cats (Mariño et al., 1999, 2000). Hence, differences may exist between cortico-cuneate interactions in cats and rats. Although pentobarbital has been reported to block cortical slow oscillations, slow rhythmicity in the

499

range of 0.1–1 Hz was consistently found in cortical neurones of cats under ketamine⫹xylazine anesthesia (Steriade et al., 1993), indicating that the absence of slow rhythms in rats cannot be ascribed to the choice of anesthesia. Our results are also in agreement with the low percentage of spontaneously rhythmic lemniscal cells reported by Nuñez et al. (2000). While we found 11% of antidromically identified spontaneous cuneolemniscal cells (7% if we take into account the silent cuneolemniscal cells) to be rhythmic, they reported 4% of low frequency gracile/cuneate cells (assumed to be lemniscal). We also agree in that rhythmicity is mainly found in cells that fire at more than 10 Hz (range: 11.3– 64.4 Hz). However, we did not find all putative interneurones (high frequency) to be rhythmic, actually in our hands, only 15% of putative interneurones (non-responding cells and those that only respond orthodromically to lemniscal stimulation) were rhythmic. Furthermore, a novel and important observation reported here is the presence of rhythmic bursting cells in the rat CN, a fact that was reported in the cat long ago. Cells firing rhythmic bursts are not affected by lemniscal stimulation (neither antidromically nor orthodromically). It is tempting to speculate that rhythmic bursting cells could be interneurones which are not related to recurrent collateral activity, and hence different from non-rhythmic bursting cells. This speculation is supported by the fact that, in long-lasting recordings, non-rhythmic cells never changed to a rhythmic firing and vice versa. Although rhythmic bursting cells always lose their rhythm when peripherally activated, this phenomenon was only temporal and they quickly recovered rhythmicity when receptive field stimulation was set off; besides, none of them responded to air-jets. On the other hand, the response of non-rhythmic bursting cells to that stimulation was always a single burst and five of them responded to air-jets. Comparison with “in vitro” experiments The range of rhythmic frequencies was essentially found to be the same in single-spiking and bursting cells (mainly 10 –30 Hz). Regarding bursting behavior, it is noteworthy that DCN cells in culture display a spontaneous rhythmic clustering behavior at 1 Hz (groups of two to five action potentials). In culture, when the membrane was depolarized, the number of action potentials by cluster increased, but the frequency remained unchanged (Reboreda et al., 2003). Note that culture experiments were carried out at room temperature so that the frequencies may be not directly comparable. On the other hand, cultured DCN cells also showed spontaneous rhythmic single-spiking activity at 14 Hz and rhythmic subthreshold membrane potential oscillations at 11 Hz (Reboreda et al., 2003). Spontaneous single-spiking rhythmicity at about 11 Hz has also been described in cuneolemniscal neurones in rat slices (Nuñez and Buño, 1999). It is tempting to speculate, although with reservations, that subthreshold oscillations observed in culture may be, at least in part, the origin of the single-spiking rhythms observed in slices and in anesthetized animals. On the other hand, the clustering behavior observed in culture could resemble the burst behavior reported here. To clarify

500

E. Sánchez et al. / Neuroscience 141 (2006) 487–500

these speculations, intracellular recording in anesthetized rat cuneate neurones should be performed.

CONCLUSION In summary, we reported three new properties of rat cuneate neurones that may be important to understand the role of this nucleus in the transmission of somatosensorial information. First, many cuneate neurones showed spontaneous bursting activity with a high intraburst frequency; second, some of them fired rhythmic bursts; and third, groups of cuneolemniscal cells which are close together showed the same conduction velocity. These characteristics of cuneate cells should improve both temporal and spatial summation at the level of cuneothalamic synapses. In turn, summation would increase their synaptic impact when they are recruited together by stimulating their overlapping peripheral receptive fields. Rhythmicity could extend these phenomena, since once initiated (locked in phase), synchronization can outlast the duration of the stimulus if the cells recruited continue to fire rhythmically and at the same frequency. Acknowledgments—This work was supported by a grant from the Spanish ministry of Science and Technology, MCYT (BFI2002-01648). E.S., A.R. and M.R. were PhD students under the Galician Predoctoral, Spanish FPI and Spanish FPU programs, respectively. We wish to thank Professor Antonio Canedo for his continuous support and criticism, and Ms. Ana Senra for her technical assistance.

REFERENCES Aguilar J, Soto C, Rivadulla C, Canedo A (2002) The lemniscalcuneate recurrent excitation is suppressed by strychnine and enhanced by GABAA antagonists in the anaesthetized cat. Eur J Neurosci 16:1697–1704. Alloway KD, Wallace MB, Johnson MJ (1994) Cross-correlation analysis of cuneothalamic interactions in the rat somatosensory system: influence of receptive field topography and comparisons with thalamocortical interactions. J Neurophysiol 72:1949 –1972. Amassian VE, Giblin D (1974) Periodic components in steady-state activity of cuneate neurones and their possible role in sensory coding. J Physiol 243:353–385. Andersen P, Eccles JC, Schmidt RF, Yokota T (1964) Identification of relay cells and interneurons in the cuneate nucleus. J Neurophysiol 27:1080 –1095. Barbaresi P, Spreafico R, Frassoni C, Rustioni A (1986) GABAergic neurons are present in the dorsal column nuclei but not in the ventroposterior complex of rats. Brain Res 382:305–326. Basbaum AI, Hand PJ (1973) Projections of cervicothoracic dorsal roots to the cuneate nucleus of the rat, with observations on cellular “bricks”. J Comp Neurol 148:347–360. Brown AG, Gordon G, Kay RH (1974) A study of single axons in the cat’s medial lemniscus. J Physiol 236:225–246. Canedo A (1997) Primary motor cortex influences on the descending and ascending systems. Prog Neurobiol 51:287–335. Canedo A, Lamas JA (1989) Rubrospinal tract of the cat: superposition of antidromic responses and changes in axonal excitability following orthodromic activity. Brain Res 502:28 –38. Canedo A, Lamas JA (1993) Pyramidal and corticospinal synaptic effects over reticulospinal neurones in the cat. J Physiol 463:475–489. Canedo A, Martínez L, Mariño J (1998) Tonic and bursting activity in the cuneate nucleus of the chloralose-anesthetized cat. Neuroscience 84:603– 617.

Canedo A, Mariño J, Aguilar J (2000) Lemniscal recurrent and transcortical influences on cuneate neurons. Neuroscience 97:317–334. Canedo A, Towe AL (1985) Superposition of antidromic responses in pyramidal tract cell clusters. Exp Neurol 89:645– 658. Davidson N, Smith CA (1972) A recurrent collateral pathway for presynaptic inhibition in the rat cuneate nucleus. Brain Res 44:63–71. Galindo A, Krnjevic K, Schwartz S (1968) Patterns of firing in cuneate neurones and some effects of Flaxedil. Exp Brain Res 5:87–101. Kemplay S, Webster KE (1989) A quantitative study of the projections of the gracile, cuneate and trigeminal nuclei and of the medullary reticular formation to the thalamus in the rat. Neuroscience 32:153–167. Lamas JA, Martínez L, Canedo A (1994) Pericruciate fibres to the red nucleus and to the medial bulbar reticular formation. Neuroscience 62:115–124. Mariño J, Aguilar J, Canedo A (1999) Cortico-subcortical synchronization in the chloralose-anesthetized cat. Neuroscience 93:409–411. Mariño J, Canedo A, Aguilar J (2000) Sensorimotor cortical influences on cuneate nucleus rhythmic activity in the anaesthetized cat. Neuroscience 95:657– 673. Mariño J, Martínez L, Canedo A (1996) Coupled slow and delta oscillations between cuneothalamic and thalamocortical neurons in the chloralose anesthetized cat. Neurosci Lett 219:107–110. Martínez L, Lamas JA, Canedo A (1995) Pyramidal and corticospinal neurons with branching axons to dorsal column nuclei of the cat. Neuroscience 68:195–206. Maslany S, Crockett DP, Egger D (1991) Somatotopic organization of the dorsal column nuclei in the rat: transganglionic labelling with B-HRP and WAG-HRP. Brain Res 564:56 – 65. Nuñez A, Buño W (1999) In vitro electrophysiological properties of rat dorsal column nuclei neurons. Eur J Neurosci 11:1865–1876. Nuñez A, Panetsos F, Avendaño C (2000) Rhythmic neuronal interactions and synchronization in the rat dorsal column nuclei. Neuroscience 100:599 – 609. Panetsos F, Nuñez A, Avendaño C (1997) Electrophysiological effects of temporary deafferentation on two characterized cell types in the nucleus gracilis of the rat. Eur J Neurosci 9:563–572. Panetsos F, Nuñez A, Avendaño C (1998) Sensory information processing in the dorsal column nuclei by neuronal oscillators. Neuroscience 84:635– 639. Paxinos G, Watson C (1998) The rat brain in stereotaxic coordinates. San Diego, CA: Academic Press. Pinault D, Deschenes M (1992) The origin of rhythmic fast subthreshold depolarizations in thalamic relay cells of rats under urethane anaesthesia. Brain Res 595:295–300. Popratiloff A, Rustioni A, Weinberg RJ (1997) Heterogeneity of AMPA receptors in the dorsal column nuclei of the rat. Brain Res 754: 333–339. Popratiloff A, Valtschanoff JG, Rustioni A, Weinberg RJ (1996) Colocalization of GABA and glycine in the rat dorsal column nuclei. Brain Res 706:308 –312. Pubols BHJ, Haring JH, Rowinski MJ (1989) Patterns of resting discharge in neurons of the raccoon main cuneate nucleus. J Neurophysiol 61:1131–1141. Reboreda A, Sánchez E, Romero M, Lamas JA (2003) Intrinsic spontaneous activity an subthreshold oscillation in neurons of the rat dorsal column nuclei in culture. J Physiol 551:191–205. Sánchez E, Reboreda A, Romero M, Lamas JA (2004) Spontaneous rhythmic activity of cuneolemniscal neurons in the rat cuneate nucleus “in vivo.” FENS Forum Abstr A222.24, P180. Steriade M, Nuñez A, Amzica F (1993) A novel slow (⬍1 Hz) oscillation of neocortical neurons in vivo: depolarizing and hyperpolarizing components. J Neurosci 13:3252–3265. Towe AL, Jabbur SJ (1961) Cortical inhibition of neurons in dorsal column nuclei of cat. J Neurophysiol 24:488 – 498.

(Accepted 27 March 2006) (Available online 3 May 2006)