| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Department of Biology and 2Department of Cellular and Molecular Medicine, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada
Submitted 2 September 2003; accepted in final form 30 October 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
). Presynaptically, there is short-term plasticity, with time scales of milliseconds (Dittman et al. 2000), as well as long-term potentiation (LTP), with time scales of hours (Salin et al. 1996). The presynaptic LTP is mediated by a cAMP/PKA-dependent pathway, and results in changes to paired-pulse facilitation that are consistent with an increase in the baseline probability of transmitter release (Castillo et al. 2002; Chen and Regehr 1997; Jacoby et al. 2001; Salin et al. 1996). Release probability also plays an important role in short-term plasticity, for instance by influencing the balance between facilitation and depression (Dittman et al. 2000). Thus there is the potential for interaction between different mechanisms of plasticity at very different time scales, but the functional implications of such interactions are not clear. To begin to understand the role of parallel fiber synapses in processing dynamic signals, a quantitative description of these interactions is necessary.
In the weakly electric fish, cerebellar granule cells provide parallel fiber inputs to the electrosensory lateral line lobe (ELL). The ELL is the first way station in the neural pathways involved in the electric sense of these fish–a sense that is critical for prey capture and conspecific communication (Heiligenberg 1991). The parallel fibers provide feedback input via synapses onto apical dendritic spines of ELL pyramidal neurons (Maler et al. 1981). These synapses exhibit a number of overlapping forms of synaptic plasticity over time scales of milliseconds to hours (Bastian 1998; Han et al. 2000; Lewis and Maler 2002) and are necessary for effective electrosensory function (Bastian 1986). Further, they express adenylate cyclase and PKA (Maler 1999) and are thus similar to mammalian cerebellar parallel fibers at many different levels.
In this study, we investigate the effect of long-term plasticity at the ELL parallel fiber synapse, lasting tens of minutes, on its short-term dynamics lasting tens of milliseconds. We first demonstrate a long-term enhancement (LTE) of this synapse that is similar to the cAMP/PKA-dependent presynaptic LTP observed in the mammalian cerebellum (Salin et al. 1996). The novelty in our approach relates to the stimulation protocol used to induce LTE. We use multiple repetitions of a random stimulus train to induce LTE and then fit a simple synaptic model (Lewis and Maler 2002) to the evoked responses. This method provides a characterization of short-term synaptic dynamics during the induction of LTE that is more accurate than that achieved using standard paired-pulse stimulation protocols. Our results show that a progressive change in release probability can explain the experimentally observed changes in short-term dynamics during LTE induction. Further, pharmacological conditions that occluded LTE also blocked the progressive change in our estimate of release probability, thus confirming the reliability of our analyses. An analysis of our model over a range of input frequencies suggests that changing the baseline probability of release, and thus the steady-state gain of the synapse, also provides a means of systematically changing the temporal filtering properties of the parallel fiber synapse.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
The gymnotiform fish Apteronotus leptorhynchus (male or female, 10–15 cm in length) were anesthetized in oxygenated water containing 0.2% 3-aminobenzoic ethyl ester (MS-222, Sigma). Surgical procedures and slice preparation were performed as previously described (Berman and Maler 1998; Lewis and Maler 2002). Briefly, true-transverse 350-µm slices of the ELL were obtained and transferred to an interface-type slice chamber. Slices were perfused (2 ml/min) with bubbled (95% O2-5% CO2), room-temperature (20–22°C), artificial cerebrospinal fluid (ACSF) containing (in mM) 124 NaCl, 24 NaHCO3, 10 D-glucose, 1.25 KH2PO4, 2 KCl, 2 CaCl2, and 2 MgSO4. A recovery period of 1–2 h was allowed before recordings were made. All pharmacological agents were bath applied. Stock solutions of Forskolin (FSK) and H-89 (Sigma) were prepared in DMSO and diluted in ACSF (to 10 and 0.2 µM, respectively) such that the final concentration of DMSO was <0.1%. All protocols were approved by the University of Ottawa Animal Care Committee.
Electrophysiological methods
Stimulation and recording methods were similar to those described previously (Lewis and Maler 2002). Parallel fibers in the ELL dorsal molecular layer (DML) were stimulated through a stimulus isolation unit (Digitimer Ltd; 100-µs pulses, 10–25 V) using two glass micro-electrodes (tip size approximately 10 µm) in a bipolar configuration, oriented perpendicularly to the parallel fibers, about 100 µm apart. Field potential recordings of the synaptic responses to stimulus trains were made using glass microelectrodes (5–10 M
filled with 1 M NaCl). The electrode tips were placed in the DML of the centromedial segment of the ELL at a depth of 30–60 µm and along the parallel fiber axis, medial to the stimulation site. Recordings were amplified and filtered (DC-1 kHz, Axoclamp-2A, Axon Instruments), digitized at 5 kHz (ITC-16, Instrutech Corporation, Greatneck, NY) and acquired using Pulse Control (Instrutech Corporation) and Igor Pro (Wavemetrics, Lake Oswego, OR) software running on a Power Macintosh 7100. We quantified the synaptic field potentials by measuring the peak deflection from baseline. The baseline was taken as the average potential in a 2-ms window just prior to stimulus onset.
The stimulation protocol (5-train protocol) used in our study consisted of five repeats of the same 200-pulse train with 5 s between each repeat. The 200 stimulus pulses within the train were separated in time by randomly distributed intervals (Poisson stimulation) with a mean interval of 62.5 ms (16 Hz mean frequency). The random intervals were computer-generated from an exponential distribution that was truncated to have a minimum interval of 10 ms (intervals smaller than this interfered with the measurement of the synaptic response). In the short term (<2 s), a mean stimulation frequency of 16 Hz results in minimal recruitment of the disynaptic inhibitory component of this feedback pathway (Lewis and Maler 2002). For 5 min before and 5 min after each train protocol, paired-pulse test stimuli were delivered every 30 s with an interpulse interval of 40 ms (for a total of 10 paired-pulse trials before and after train stimulation). The measure of synaptic efficacy was taken from the response to the first of this pair of pulses. The data reported for the test stimuli are the averages of two trials during a given minute. Paired-pulse facilitation (PPF) was calculated as the amplitude of the second response divided by that of the first and was used for comparison with previous studies of cerebellar plasticity (Salin et al. 1996). Data from individual slices were from a single presentation of the stimulus protocol, except in five slices, in which data were an average of two presentations with 
30 min between. All data are reported as mean ± SE in either absolute or normalized units (unless otherwise noted). Statistical comparisons among data were performed using an ANOVA or a paired t-test.
The stimulation methods we use can sometimes be associated with changes in the recruitment of the stimulated fibers, reflected in changes in the presynaptic fiber volley (Salin et al. 1996). In some of our experiments, the fiber volley was difficult to distinguish from the stimulus artifact, but we were able to do so in nine separate experiments. In these cases, the peak negativity associated with the fiber volley was slightly decreased in the 2- to 5-min posttrain period (normalized to pretrain value, 0.91 ± 0.048, P = 0.01), but in no cases did we observe evidence of an increased fiber volley. In rat cerebellar studies, there is also an initial decrease in the fiber volley immediately following the stimulus period, followed by a slight increase (Salin et al. 1996).
Model description
We have previously described ELL parallel fiber short-term plasticity on time scales of less than a second using the so-called FD formalism (Dayan and Abbott 2001; Lewis and Maler 2002). In an FD model, a stimulus-evoked postsynaptic potential (PSP) is described by a combination of facilitation-like processes (F) and depression-like processes (D) that increase or decrease, respectively, with the PSP amplitude relative to the control amplitude (Ao). This general formalism has been used to describe many different synapses with the details of implementation varying among these cases (see references in Lewis and Maler 2002; Zucker and Regehr 2002). Our original model was based on a previous FD model of a cerebellar parallel fiber synapse (Dittman et al. 2000) to which we added a depression-like term describing disynaptic inhibition. This inhibition, which is due to parallel fiber activation of interneurons in the ELL molecular layer, is activated at mean stimulation frequencies > 16Hz; this is true for both fixed-interval and random-interval stimulation (Lewis and Maler 2002), suggesting that a sequence of multiple high-frequency events is required to recruit inhibition. In this paper, we present a slightly simplified version of our original model (Lewis and Maler 2002) to facilitate our analyses. First, because we limit our stimulation frequency to a mean of 16 Hz, we can discount the inhibitory process. Second, the F process in the original model contained a nonlinear squashing function (as in Dittman et al. 2000; see Eq. 3 in Lewis and Maler 2002) so that F would not increase beyond a value of 1. Here, we eliminate this nonlinearity because the squashing function is nearly linear for the relatively low stimulation frequencies considered. Our reduced model is summarized in the next set of equations (Eqs. 1, 2, 3)
 | (1) |
 | (2) |
 | (3) |
F and then recovers exponentially to its resting value (F0) with a time constant 
F. Similarly, at each stimulus, D is decreased by an amount given by the product FD and recovers exponentially to a value of 1 with a time constant D. Both F and D terms can be associated with the calcium-dependent vesicle release process, with the product FD indicating transmitter release probability and F0 as the baseline release probability (Dittman et al. 2000).
In summary, the present model has five parameters (A0, F0, F, F, and D). The parameter D is fixed to 0.083 s, as in our previous study (Lewis and Maler 2002); A0 is a constant scaling factor set to 1/F0 when we consider normalized responses. The remaining parameters (F0, F, and F) are considered free parameters and will be addressed individually under RESULTS.
Model simulations
In Figs. 4 and 5, we show results from model simulations, using different stimulation frequencies and different values of the parameter F0 in each simulation. We delivered random Poisson trains (stimulus intervals ti that are exponentially distributed as in the experiments) consisting of 1000 pulses to our model at mean frequencies from 1 to 50 Hz. For each mean frequency, we computed the steady-state response over the train in terms of the mean normalized PSP size given by the model (Eq. 1 with A0 = 1/F0). In general, there was a particular frequency for which the response was maximal (see dotted lines in Fig. 5A; this was designated as the preferred frequency and is the quantity plotted in Fig. 5B as a function of F0. We also calculated the SD and coefficient of variation (CV = SD/mean) of these responses for one stimulation frequency (16 Hz) to provide the predictions shown in Fig. 4B.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
The goal of the present study was to characterize the changes in short-term synaptic plasticity (time scales of tens of milliseconds) during the development of synaptic changes on much longer time scales (tens of minutes). We first describe a LTE produced by patterned stimulation of ELL parallel fibers that is similar to the presynaptic LTP expressed by parallel fibers in the rat cerebellum (Castillo et al. 2002; Chen and Regehr 1997; Jacoby et al. 2001; Salin et al. 1996).
To induce LTE, we used a stimulus protocol lasting slightly <90 s and consisting of five repetitions of a 200-pulse 16-Hz Poisson train with 5 s between each train (5-train protocol, see METHODS). This protocol resulted in an enhancement of the field potential excitatory postsynaptic potential (fEPSP) amplitude by about 30% (n = 12 slices; Fig. 1A). We monitored this enhancement at regular intervals for 5 min after train stimulation. LTE lasted about 25 min, at which time the normalized fEPSP was 1.06 ± 0.054 and not significantly different from control (P = 0.34), but we did not systematically track the entire time course of recovery. We refer to this phenomenon as a LTE. PPF was variable, but in parallel with the LTE, there was a significant decrease (15%; P = 0.02; n = 12 slices) in PPF in the posttrain period (2–5 min): PPF was 2.06 ± 0.13 in the pretrain control period and 1.78 ± 0.078 in the posttrain period. At 25 min posttrain, PPF increased to 1.87 ± 0.17 and was not significantly different from pretrain controls (P = 0.30). LTE did not occur with a single repetition of the 16-Hz Poisson train (n = 8 slices; Fig. 1A); during the posttrain period following a single train (2–5 min) the fEPSP amplitudes, as well as PPF (2.07 ± 0.15), were not significantly different from control (P = 0.80 and P = 0.31, respectively).
|
), we applied the adenylate cyclase activator FSK (10 µM) to the bath. FSK induced an enhancement (+50%) similar to that described in rat cerebellar slices (n = 5 slices). The mean normalized fEPSP increased after FSK treatment to 1.53 ± 0.13 (P = 0.001 compared with control) while PPF decreased to 1.6 ± 0.13 (P = 0.03). The 5-train protocol delivered during the FSK-induced enhancement did not produce further enhancement of the normalized fEPSP amplitude (Fig. 1B; P = 0.47 compared with pretrain) nor did it change PPF (1.50 ± 0.14; P = 0.15). Pretreatment with FSK therefore occludes the train-induced LTE that we observe in ELL. Further, there was no significant LTE when we repeated the 5-train protocol in the presence of the PKA inhibitor H-89 (0.2 µM; n = 3); under these conditions there was a slight but insignificant enhancement (P = 0.07; normalized posttrain fEPSP amplitude of 1.183 ± 0.021). Overall, these data suggest that the LTE is mediated by a presynaptic cAMP/PKA-dependent signaling pathway that is biochemically similar to the previously described cerebellar LTP.
A number of overlapping processes are thought to be involved in long-term plasticity at the ELL parallel fiber synapse (Bastian 1999; Han et al. 2000), among them a postsynaptic N-methyl-D-aspartate (NMDA)-dependent long-term depression (LTD). To determine if this LTD was activated by our stimulus protocol and was thus acting to offset the expression of LTE, we repeated the 5-train protocol in the presence of the NMDA antagonist 2-amino-5-phosphonovaleric acid (APV; 50 µM). With APV, there was a significant increase in the level of LTE, as compared with that observed under control conditions, only during the first minute of the posttrain period (n = 5; P = 0.03) (see Fig. 1, open symbols). Later in the posttrain period (2–5 min), the increase in the level of LTE became insignificant (P = 0.18). In contrast, during the entire posttrain period, PPF was not significantly different between control and APV conditions (1 min posttrain, P = 0.7; 2–5 min posttrain, P = 0.6). These observations suggest that the transient depression evident in the 1-train and FSK conditions (Fig. 1) may be postsynaptically mediated and NMDA dependent. Thus this depression could be acting to control the expression of the LTE we have described (we did not pursue this possibility further in the current study).
Short-term synaptic dynamics during LTE induction
Because our stimulation protocol consisted of multiple repeats of a single Poisson train, we were able to efficiently assess the short-term dynamics of this synapse over a range of frequencies. This marks an improvement over previous studies of long-term plasticity, using paired-pulse protocols and fixed-interval trains that involve only one interstimulus interval, and thus sample the synaptic dynamics at only one frequency. Here, we considered the responses to the first 32 stimuli of each random train and used our model to quantify the changes in these dynamics during the onset of LTE. We limited our analysis to this brief subset of responses to preclude significant influences of long-term processes; because long-term plasticity involves slow time scales, their influence changes relatively little over the time scales that short-term processes act. Further, the 5 s between successive trains allowed any short-term dynamics (time constants less than 1 s) to recover. In addition to our pharmacological comparison with presynaptic LTP in rat cerebellum (Salin et al. 1996), we present here a novel method of testing whether the changes in synaptic dynamics during the onset of LTE in ELL could also be explained by changes in the baseline probability of release. Baseline probability of release is directly, and uniquely, related to the parameter F0 in our model (Dittman et al. 2000) and thus was expected to increase during LTE.
Our model involves three free parameters (F0, F, and F). So to test the impact of varying F0 alone, we first fit the model to the mean responses (n = 12) over the first 32 stimuli of the first train in the 5-train protocol (F0 = 0.10; F = 0.23; F = 0.079 s; 10.4% error; Fig. 2, top). The parameters F and F then were held constant at these values, with F0 as the only free parameter. Figure 2 shows the mean response data (normalized to the first response of the first train) over the first 2 s of stimulation in the 1st, 3rd, and 5th trains, along with the model fits (and corresponding estimates of F0). With F0 as the only free parameter, the model accurately fit the mean data for each of the five trains, as indicated by the small fitting errors (Fig. 3B, open circles). In particular, the model captures two prominent aspects of the data: the decrease in response variance over successive trains and the decrease in peak response. We emphasize that the pattern of stimulation in each train is identical, so that the differences observed over the trains (other than those due to random variations) are due to longer-term changes in the dynamics of the parallel fiber synapse. Performing a similar analyses with either F or F as the free parameters (and the other two parameters held constant) resulted in much higher fitting errors (Fig. 3B), whereas leaving all three parameters free to vary led to only a marginal decrease in the fitting error compared with varying F0 alone (Fig. 3B, closed squares). This suggests that the changing synaptic dynamics over our stimulation protocol can be explained largely by changes in the parameter F0, suggesting that the transmitter release probability is increasing during the onset of the LTE.
|
|
In the previous section, we suggested that treating slices with FSK occludes the train-induced LTE by increasing the probability of transmitter release via a similar mechanism. This provides a specific prediction about how the parameter F0 should vary during train stimulation in the presence of FSK. First, the value of F0 should be greater than control during the 1st train. Second, if the LTE is occluded by FSK and not simply masked, then F0 should not change during the 5-train protocol. This is exactly what we observed (Fig. 3A, closed squares).
As mentioned previously, response variability decreases over the course of the 5 random trains (Fig. 2). To quantify this behavior we calculated the SDs of the responses over the first 2 s of stimulation for each of the 5 trains. We used the mean response data so that any fluctuations due to noise (i.e., not due to the dynamics per se) would be averaged out. Under control conditions, there is a clear decrease in SD during the 5-train protocol, whereas during FSK treatment, the SD is lower and does not show a trend with train number (Fig. 4A). This result parallels that shown in Fig. 3A and is due mainly to the changing release probability (model parameter F0). This is illustrated by plotting the SD as a function of the value of F0 found for the corresponding train (Fig. 4B; open circles, control; open squares, FSK). The solid line in Fig. 4B shows the SD of the model responses 16-Hz random stimulation over the same range of F0. The data for both control and FSK conditions clearly follow the model predictions, suggesting that a main determinant of response variability under these conditions is the baseline release probability. Also shown in Fig. 4B is a similar relationship for the CV, verifying that there are no major effects due to a change in mean response amplitude.
Taken all together, our results show that the dynamics of the parallel fiber synapse in ELL during LTE onset can be accounted for by changes in the parameter F0 alone, illustrating in a novel way that the observed LTE is attributable to factors influencing the baseline probability of transmitter release.
Frequency dependence of short-term synaptic dynamics
We developed our model under simplifying conditions to facilitate our analysis. Nonetheless, we can probe the simplified model under more general conditions to gain insight into the functional implications of changing baseline release probability (Dittman et al. 2000), for example, via mechanisms such as the LTE we describe in this study.
In a set of simulations, we delivered random Poisson trains (1000 pulses) to our model using mean frequencies from 1 to 50 Hz. For each frequency of random stimulation and choice of value for F0, the mean normalized PSP size (Eq. 1; A0 = 1/F0) over the stimulation period was calculated, thus providing the frequency response for the model. Figure 5A shows the results for two different values of F0 (0.1 and 0.2). Similar to what we have previously described for the more complicated model (Lewis and Maler 2002), the present model exhibits a bandpass behavior; the PSP size is maximal for a particular mean frequency of stimulation, preferred frequency (dotted vertical lines in Fig. 5A). Interestingly, this preferred frequency decreases with increasing values of the parameter F0 in an almost linear manner (Fig. 5B). The details of this behavior will depend on the parameters associated with both F and D, but we have used the parameter values that best fit our data. The CV of the responses at this preferred frequency decreases in a similar way with F0 (not shown, but see Fig. 4B for 16-Hz stimulation), suggesting that a number of synaptic filtering properties can be influenced and controlled by processes such as the LTE we describe in this paper. In this way, long-term synaptic enhancement can be thought of, not only as a mechanism for memory storage, but also as a means for changing synaptic filtering on much shorter time scales.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
We have demonstrated in this paper that the parallel fiber feedback pathway to the ELL of an electric fish exhibits a LTE that is similar to the presynaptic LTP observed in rat cerebellum (Salin et al. 1996). Further, we have presented evidence that, during the induction of LTE, there is a progressive increase in the probability of transmitter release at these ELL synapses. In addition, our study builds on previous work in two important ways.
First, we provided an accurate characterization of short-term dynamics at the parallel fiber synapse over a range of frequencies. Our methods follow from work in other systems that have used Poisson trains to characterize short-term synaptic dynamics (Markram and Tsodyks 1996; Varela et al. 1997). However, previous work on long-term enhancement of the parallel fiber synapse in mammalian cerebellum has involved fixed-interval trains and paired-pulse stimulation protocols, which sample only a single frequency at a time (Salin et al. 1996). While these studies have also linked changes in release probability to LTE at the parallel fiber synapse, none have used methods that quantified the dependence of these changes over a range of frequencies.
Second, we have characterized the evolution of short-term synaptic dynamics during the induction of LTE and not simply after its expression. This is an important distinction because, under normal physiological conditions, there is likely a dynamic balance between many different types of plasticity. In such situations, an understanding of the slow changes in synaptic plasticity will be critical. A few previous studies of other synapses have monitored short-term synaptic dynamics before and after the induction of long-term synaptic plasticity (Finnerty and Connors 2000; Finnerty et al. 1999; Markram and Tsodyks 1996; Toth et al. 2000). In the cases in which changes in short-term dynamics were observed, they were also consistent with a change in the baseline probability of transmitter release (Finnerty et al. 1999; Markram and Tsodyks 1996; Toth et al. 2000). However, because of the methods used, none of these studies were able to monitor synaptic dynamics while the long-term plasticity was developing.
Under natural circumstances, the parallel fiber synapses in ELL will be influenced by a number of dynamic processes (expressed both pre- and postsynaptically), overlapping on a variety of time scales. For simplicity, we have limited our studies to two specific forms of presynaptic plasticity: short-term plasticity (facilitation and depression) with a time scale near 100 ms and a LTE with a time scale around 20 min. There are no guarantees that postsynaptic changes are not influencing our results. However, it is generally accepted that the parallel fiber LTP in mammalian cerebellum is presynaptically mediated. It is likely, by analogy, that the LTE we describe is presynaptic as well. More important though is that long-term postsynaptic changes would have to affect short-term dynamics in a manner that is very similar to those resulting from changes in baseline release probability. The fact that our model, based on short-term facilitation and depression, can explain the changes in dynamics during LTE is further evidence for a presynaptic locus. Postsynaptic modifications (e.g., changes in receptor desensitization) may influence short-term dynamics (Zucker and Regehr 2002), but long-term postsynaptic changes are usually associated with processes that change the gain of the synapse without affecting the short-term dynamics.
In the least, our stimulation protocol does recruit a depression-like process with a time scale of a few minutes (Fig. 1), but we do not yet know its underlying mechanism. It may be related to the postsynaptic NMDA-dependent LTD expressed at this synapse (Bastian 1998); our stimulation protocol is much different in pattern, frequency, and duration from that used by Bastian (1998), so we may be observing a partial activation of this LTD. Nonetheless, such an LTD could be acting to offset the LTE we observe so that the synaptic gain is tightly regulated. In this way, the dynamics (in terms of short-term filtering) could be regulated by LTE independent of any net change in synaptic gain.
Another process contributing to the dynamics at the ELL parallel fiber synapse, although not directly related to synaptic plasticity, is inhibition. We have previously described this disynaptic inhibition (Lewis and Maler 2002), but for simplicity have purposely chosen a stimulation protocol in the present study that minimizes its influence. However, under natural conditions, this inhibition will certainly lead to more diverse dynamics, by controlling both synaptic gain and filtering properties.
In general, the functional implications of long-term changes in release probability on short-term synaptic dynamics are not known but have been extensively hypothesized (Tsodyks and Markram 1997). Short-term plasticity confers specific filtering properties to a synapse (Fortune and Rose 2001), and thus changing these dynamics through long-term plasticity will lead to different filtering properties (Fig. 5) (Abbott et al. 1997; Buonomano 2000; Tsodyks and Markram 1997). Some forms of long-term plasticity may then provide a means of controlling both gain and frequency selectivity of a neuronal pathway. In this context, the ELL parallel fiber system of the weakly electric fish provides an ideal opportunity to test these hypotheses. These synapses exhibit long-term changes that are involved in sensory processing on time scales greater than tens of minutes (Bastian 1998; Han et al. 2000), and yet they also exhibit short-term dynamics (time scales of seconds) that are particularly relevant for prey detection (Lewis and Maler 2002; MacIver et al. 2001). Because of their proximity to the sensory periphery, as well as their strong link to a functional context, these synapses provide an excellent system by which to study the role of synaptic plasticity on multiple time scales in adaptive sensory processing (Bastian 1999; Bell 2001).
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
GRANTS
This work was supported by a Canadian Institutes of Health Research (CIHR) Senior Research Fellowship to J. E. Lewis and a CIHR operating grant to L. Maler.
|
FOOTNOTES |
|---|
Address for reprint requests: J. E. Lewis, Department of Biology, University of Ottawa, 150 Louis-Pasteur, Ottawa, ON, K1N 6N5 Canada (E-mail: jlewis{at}uottawa.ca).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Bastian J. Gain control in the electrosensory system: a role for the descending projections to the electrosensory lateral line lobe. J Comp Physiol A 158: 505-515, 1986.[CrossRef][Medline]
Bastian J. Plasticity in an electrosensory system. III. Contrasting properties of spatially segregated dendritic inputs. J Neurophysiol 79: 1839-1857, 1998.
Bastian J. Plasticity of feedback inputs in the apteronotid electrosensory system. J Exp Biol 202(Pt 10): 1327-1337, 1999.[Abstract]
Bell CC. Memory-based expectations in electrosensory systems. Curr Opin Neurobiol 11: 481-487, 2001.[CrossRef][ISI][Medline]
Berman NJ and Maler L. Inhibition evoked from primary afferents in the electrosensory lateral line lobe of the weakly electric fish (Apteronotus leptorhynchus). J Neurophysiol 80: 3173-3196, 1998.
Buonomano DV. Decoding temporal information: a model based on short-term synaptic plasticity. J Neurosci 20: 1129-1141, 2000.
Castillo PE, Schoch S, Schmitz F, Sudhof TC, and Malenka RC. RIM1alpha is required for presynaptic long-term potentiation. Nature 415: 327-330, 2002.[CrossRef][Medline]
Chen C and Regehr WG. The mechanism of cAMP-mediated enhancement at a cerebellar synapse. J Neurosci 17: 8687-8694, 1997.
Dayan P and Abbott LF. Theoretical Neuroscience. Cambridge, MA: MIT Press, 2001.
Dittman JS, Kreitzer AC, and Regehr WG. Interplay between facilitation, depression, and residual calcium at three presynaptic terminals. J Neurosci 20: 1374-1385, 2000.
Finnerty GT and Connors BW. Sensory deprivation without competition yields modest alterations of short-term synaptic dynamics. Proc Natl Acad Sci USA 97: 12864-12868, 2000.
Finnerty GT, Roberts LS, and Connors BW. Sensory experience modifies the short-term dynamics of neocortical synapses. Nature 400: 367-371, 1999.[CrossRef][Medline]
Fortune ES and Rose GJ. Short-term synaptic plasticity as a temporal filter. Trends Neurosci 24: 381-385, 2001.[CrossRef][ISI][Medline]
Han VZ, Grant K, and Bell CC. Reversible associative depression and nonassociative potentiation at a parallel fiber synapse. Neuron 27: 611-622, 2000.[CrossRef][ISI][Medline]
Hansel C, Linden DJ, and D'Angelo E. Beyond parallel fiber LTD: the diversity of synaptic and non-synaptic plasticity in the cerebellum. Nat Neurosci 4: 467-475, 2001.[ISI][Medline]
Heiligenberg W. Neural Nets in Electric Fish. Cambridge, MA: MIT Press, 1991.
Jacoby S, Sims RE, and Hartell NA. Nitric oxide is required for the induction and heterosynaptic spread of long-term potentiation in rat cerebellar slices. J Physiol 535: 825-839, 2001.
Lewis JE and Maler L. Dynamics of electrosensory feedback: short-term plasticity and inhibition in a parallel fiber pathway. J Neurophysiol 88: 1695-1706, 2002.
MacIver MA, Sharabash NM, and Nelson ME. Prey-capture behavior in gymnotid electric fish: motion analysis and effects of water conductivity. J Exp Biol 204: 543-557, 2001.[Abstract]
Maler L. Distribution of adenylate cyclase in the brain of Apteronotus leptorhynchus as revealed by forskolin binding. J Comp Neurol 408: 170-176, 1999.[CrossRef][ISI][Medline]
Maler L, Sas EK, and Rogers J. The cytology of the posterior lateral line lobe of high-frequency weakly electric fish (Gymnotidae): dendritic differentiation and synaptic specificity in a simple cortex. J Comp Neurol 195: 87-139, 1981.[CrossRef][ISI][Medline]
Markram H and Tsodyks M. Redistribution of synaptic efficacy between neocortical pyramidal neurons. Nature 382: 807-810, 1996.[CrossRef][Medline]
Salin PA, Malenka RC, and Nicoll RA. Cyclic AMP mediates a presynaptic form of LTP at cerebellar parallel fiber synapses. Neuron 16: 797-803, 1996.[CrossRef][ISI][Medline]
Toth K, Suares G, Lawrence JJ, Philips-Tansey E, and McBain CJ. Differential mechanisms of transmission at three types of mossy fiber synapse. J Neurosci 20: 8279-8289, 2000.
Tsodyks MV and Markram H. The neural code between neocortical pyramidal neurons depends on neurotransmitter release probability [published erratum appears in Proc Natl Acad Sci USA 94: 5495, 1997]. Proc Natl Acad Sci USA 94: 719-723, 1997.
Varela JA, Sen K, Gibson J, Fost J, Abbott LF, and Nelson SB. A quantitative description of short-term plasticity at excitatory synapses in layer 2/3 of rat primary visual cortex. J Neurosci 17: 7926-7940, 1997.
Zucker RS and Regehr WG. Short-term synaptic plasticity. Annu Rev Physiol 64: 355-405, 2002.[CrossRef][ISI][Medline]
This article has been cited by other articles:
|
|
 |
|
  J. E. Lewis, B. Lindner, B. Laliberte, and S. Groothuis Control of neuronal firing by dynamic parallel fiber feedback: implications for electrosensory reafference suppression J. Exp. Biol., December 15, 2007; 210(24): 4437 - 4447. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
