JN Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Neurophysiol 79: 1839-1857, 1998;
0022-3077/98 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bastian, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bastian, J.

The Journal of Neurophysiology Vol. 79 No. 4 April 1998, pp. 1839-1857
Copyright ©1998 by the American Physiological Society

Plasticity in an Electrosensory System. III. Contrasting Properties of Spatially Segregated Dendritic Inputs

J. Bastian

Department of Zoology, University of Oklahoma, Norman, Oklahoma 73019

    ABSTRACT
Abstract
Introduction
Methods
Results
Discussion
References

Bastian, J. Plasticity in an electrosensory system. III. Contrasting properties of spatially segregated dendritic inputs. J. Neurophysiol. 79: 1839-1857, 1998. Efferent neurons of the first-order electrosensory processing center of the brain, the electrosensory lateral line lobe (ELL), receive electroreceptor afferent input as well as feedback inputs descending from higher centers. These ELL efferents, pyramidal cells, adaptively filter predictable patterns of sensory input while preserving sensitivity to novel stimuli. The filter mechanism involves integration of centrally generated predictive inputs with the afferent inputs being canceled. The predictive inputs, referred to as "negative image" inputs, terminate on pyramidal cell apical dendrites and generate responses that are opposite those resulting from the predictable afference, hence integration of these signals results in attenuation of pyramidal cell responses. The system also shows a robust form of plasticity; the pyramidal cells learn, with a time course of a few minutes, to cancel new patterns of repetitive inputs. This is accomplished by adjusting the strength of excitatory and inhibitory apical dendritic inputs according to an anti-Hebbian learning rule. This study focuses on the properties of two separate pathways that convey descending information to pyramidal cell apical dendrites. One pathway terminates proximally, nearer to the pyramidal cell body, whereas the other terminates distally. Recordings of ELL evoked potentials, extracellular pyramidal cell spike responses, and intracellularly recorded synaptic potentials show that the pyramidal cells respond oppositely to moderate-frequency (> ~8 Hz) single pulse stimulation or repeated (1/s) tetanic activation of these two pathways. Repetitive activation of the proximally terminating pathway results in highly facilitating responses due to potentiation of pyramidal cell excitatory postsynaptic potentials (EPSPs). These same stimuli applied to the distally terminating pathway result in a reduction of pyramidal cell responses due to depression of EPSPs and potentiation of inhibitatory postsynaptic potentials (IPSPs). Anti-Hebbian plasticity was demonstrated by pairing tetanic stimulation of either pathway with changes in the postsynaptic cell's membrane potential. After stabilization of the response potentiation due to tetanic stimulation of the proximally terminating pathway, paired postsynaptic hyperpolarization resulted in further increases in spike responses and additional potentiation of pyramidal cell EPSPs. Paired postsynaptic depolarization reduced subsequent responses to the tetanus, depressed EPSP amplitudes, and, in many cases, potentiated IPSPs. The same pattern of plasticity was observed when postsynaptic hyper- or depolarization was paired with tetanic stimulation of the distally terminating pathway except that the plasticity was superimposed on the depressed pyramidal cell responses resulting from stimulating this pathway alone. Modulation of a postsynaptic form of synaptic depression is proposed to account for the anti-Hebbian plasticity associated with both pathways.

    INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References

Recent studies of several lower vertebrate acousticolateralis systems have demonstrated that adaptive filtering mechanisms exist that suppress responses to predictable patterns of sensory input without eliminating responses to novel stimuli. The rejection of predictable signals occurs in the initial processing station in the brain and the cancellation mechanism involves the integration of "negative image inputs" with the inputs from the periphery to be canceled (Bell et al. 1997a). This phenomenon was discovered by Bell (1981, 1982). The negative image inputs supply net inhibition to their target cells when the unwanted signals are excitatory and vice versa. The negative image inputs are specific; they cancel only the predictable component of a complex sensory input leaving novel features unaltered, and they are plastic; the negative image inputs can be modified to cancel new stimulus patterns if these become predictable.

Bell's studies of this cancellation mechanism, done with mormyrid weakly electric fish, showed that second-order cells receiving a specific category of electroreceptor input (ampullary input) were unresponsive to the animal's electric organ discharge (EOD) despite the fact that the receptors responded well. Ampullary receptors are designed to detect low-frequency signals of extrinsic origin, hence cancellation of responses to the animal's own discharge frees the system from processing this highly stereotyped signal. This is thought to increase sensitivity to more relevant stimuli (Bell 1982). The EOD activation of ampullary receptors, an example of a reafferent sensory input, is predictable because the animals "know" when the electric organ discharges. An EOD command related signal, generated by the circuitry that controls the electric organ, provides the negative image information to the cells implementing the cancellation (Bell and Grant 1992; Bell and von der Emde 1995; Bell et al. 1993, 1995).

A similar mechanism has been discovered in the electrosensory system of marine rays. In this case, reafferent electrosensory stimuli resulting from the animal's gill movements and movements of the fins are canceled at the first processing station in the brain. Proprioceptive information related to gill or fin movements, corollary discharges of motor commands to these structures, and electrosensory inputs descending from higher centers contribute to the negative image inputs (Hjelmstad et al. 1996; Montgomery and Bodznick 1994). Montgomery and Bodznick also have shown that cancellation of reafferent mechanosensory inputs due to gill movements occurs in teleost fish.

A fourth example of the cancellation of predictable inputs comes from studies of gymnotid weakly electric fish. The amplitude of EOD signal received by these animals' electroreceptors depends on body geometry or posture. Stereotyped movements, such as occur during swimming and exploration of the environment, result in modulations of the EOD field amplitude that strongly drive the electroreceptors. Pyramidal cells within the initial electrosensory processing nucleus, the electrosensory lateral line lobe (ELL) are, however, insensitive to these modulations (Bastian 1995). As in the examples described above, the rejection of this reafferent signal results from the integration of negative image inputs with those from the periphery. Both proprioceptive inputs encoding aspects of posture as well as electrosensory inputs descending from higher centers contribute to the cancellation (Bastian 1996a).

In all of these examples, the anatomy of the brain regions receiving the primary afferent input is similar and cerebellar-like (Montgomery et al. 1995). Sensory afferent inputs terminate on basilar or somatic dendrites of the structure's principal efferent neurons and the negative image inputs terminate on the cells' extensive apical dendrites. Integration of these signals results in strong attenuation of the input from the periphery. The plasticity underlying the adaptive characteristic of the cancellation is described as anti-Hebbian (Bell et al. 1993). That is, correlated increases in pre- and postsynaptic activity lead to reduced excitation provided by the negative image pathways and, in some cases, increases in the strength of inhibitory inputs. Presynaptic activity in conjunction with decreased afferent input leads to enhanced excitation provided by the negative image inputs. The anti-Hebbian nature of the plasticity ensures that the effects of the negative image inputs are updated to cancel optimally rather than reinforce predictable patterns of afferent input.

The initial electrosensory processing station of gymnotids, the ELL, is illustrated diagrammatically in Fig. 1. The basilar pyramidal cell shown is one of two major types of ELL efferent neurons. These receive direct excitatory input from electroreceptor afferents. The second category, nonbasilar pyramidal cells, receive input from inhibitory interneurons, which are driven by receptor afferent input (Maler et al. 1981). Hence, basilar pyramidal cells, referred to as E cells, are excited by increased afferent activity, whereas the nonbasilar cells (I cells) are inhibited by such stimuli (Saunders and Bastian 1984). Both types of pyramidal cell send extensive apical dendrites into the ELL molecular layers. These pyramidal cells project to two higher centers, the torus semicircularis and the nucleus praeeminentialis (nP), and this latter structure is the source of descending electrosensory inputs to the ELL molecular layers (Sas and Maler 1983, 1987).


View larger version (96K):
[in this window]
[in a new window]
  		FIG. 1. Schematic illustration of the electrosensory lateral line lobe (ELL) of Apteronotus leptorhynchus. BP, basilar pyramidal cell; EGp, eminentia granularis posterior; nP, nucleus praeeminentialis; nVML, ventral molecular layer neuron.

The ELL molecular layer consists of dorsal (lightly shaded) and ventral (heavily shaded) subdivisions and axons within each of these provide negative image inputs to the pyramidal cells. The dorsal molecular layer consists of typical cerebellar parallel fibers originating in an overlying mass of granule cells, the posterior eminentia granularis (EGp); these fibers terminate on the more distal regions of the pyramidal cell's apical dendrites. The EGp receives electrosensory inputs descending from the nP as well as proprioceptive inputs, which indicate the position of the animal's trunk and tail (Bastian 1995). Axons terminating within the ventral molecular layer also originate in the nP, and these convey descending electrosensory information directly to the proximal regions of the pyramidal cell apical dendrites. Both excitatory (glutamatergic) and inhibitory (GABAergic) inputs have been described (Maler and Mugnaini 1994; Wang and Maler 1994), and results of earlier studies verified that the ventral molecular layer inputs are plastic and can contribute to the negative image inputs (Bastian 1996a,b).

In these previous experiments, direct electrical stimulation of the ventral molecular layer inputs was paired with local electrosensory stimulation of a pyramidal cell's receptive field or changes in the cell's membrane potential due to intracellular current injection. Pairing ventral molecular layer (VML) activation with reduced electrosensory input, or with pyramidal cell hyperpolarization, potentiated subsequent VML-evoked excitatory postsynaptic potentials (EPSPs). Conversely, pairing VML stimulation with increased afferent activation of pyramidal cells, or with depolarizing current injection, resulted in the potentiation of VML-evoked inhibitatory postsynaptic potentials (IPSPs). These changes in synaptic strength indicated the presence of anti-Hebbian plasticity at the VML fiber to pyramidal cell synapses as is required to account for the neurons' ability to cancel predictable patterns of sensory input.

This study focuses on the role of the dorsal molecular layer in this cancellation mechanism and compares pyramidal cell responses evoked by activity in the distally terminating pathway with responses to activation of the proximal inputs. Previous physiological studies have described the responses of descending neurons, which give rise to the major excitatory input to the VML, and electrical stimuli were designed to mimic these cells' normal firing pattern (Bratton and Bastian 1990). Comparable information about the firing characteristics of EGp granule cells, which give rise to dorsal molecular layer (DML) parallel fibers, is not available, therefore various temporal patterns of DML and VML activation were used as well as those that mimic normal VML inputs. In addition to comparing the responses of pyramidal cells to various patterns of DML and VML activation, experiments were done to determine if DML inputs are plastic and, if so, how the DML plasticity compares with the anti-Hebbian plasticity described for the VML inputs.

    METHODS
Abstract
Introduction
Methods
Results
Discussion
References

The weakly electric fish Apteronotus leptorhynchus was used in these studies. Animals were housed in groups of 5-10 in 150 l aquaria; temperature and conductivity ranged from 25 to 27°C and 125 to 250 µS, respectively. Before surgery animals were anesthetized with tricain methanesulfonate (MS222, ~100 ppm) followed by immobilization with intramuscular injection of gallamine triethiodide (100-200 µl of 2 mg/ml solution). Fish were transferred to a 30 × 30 × 7 cm recording chamber, and the skin covering the skull was infiltrated with 2% lidocaine. After removal of the overlying skin and tissue, the rostral skull was glued to a rigid support and the ELL and cerebellum was exposed. During the course of the experiment, animals were respirated with a continuous flow of aerated water (temperature, 25-27°C; water conductivity, 100 µS).

Recording techniques

Field potential recordings were made with glass pipettes filled with 1 M NaCl (tips typically between 5 and 10 µm, electrode impedance <10 MOmega ), preamplified with a WPI DUO 773 electrometer and digitized at 20 kHz with Cambridge Electronic Design (CED) 1401+ data acquisition hardware and Spike2 software. Extracellular single unit recordings were made with metal filled glass microelectrodes (Frank and Becker 1964). Signals were recorded differentially via matched electrodes and amplified with a WPI DAM 70 preamplifier (300-Hz to 10-kHz filters), single spikes were isolated with a custom-built threshold detector, and spike times recorded and subsequently analyzed with the CED system. Intracellular recordings were made with alumino or borosilicate glass electrodes made with a P-87 Flaming/Brown electrode puller and filled with 3 M K-acetate. Before use electrodes were beveled (K. T. Brown BV-10 beveler, 0.05-µm alumina grinding disk) to reduce impedance to 100-150 MOmega . Electrodes were advanced with a Burleigh piezoelectric microdrive. Judging from action potential waveforms, dendritic as well as somatic recordings were made (Turner et al. 1994), and only cells having stable resting potentials >50 mV were studied. In both extracellular and intracellular recording experiments, neurons were identified as basilar or nonbasilar pyramidal cells on the basis of stereotyped responses to stepwise changes in electric organ discharge amplitude (Saunders and Bastian 1984). Neurons were not intracellularly labeled in this study, however, based on locations of electrode penetrations, it is likely that most recordings were made from the centrolateral and centromedial subdivisions of the ELL. Intracellular records were analog tape recorded (DC-1.25 kHz bandwidth) and later A to D converted, 5-kHz sampling rate, and analyzed via the CED hardware and software.

Stimulation techniques

Electrosensory stimuli consisted of stepwise increases or decreases in the animal's EOD amplitude. Single cycles of a sinusoidal signal, having a period slightly less than that of the EOD, were synchronized to the discharge and applied as 100-ms bursts either in phase or in antiphase relative to the animal's discharge waveform to generate increases or decreases in EOD amplitude, respectively. This signal was isolated from ground via a WPI A395 linear stimulus isolation unit and applied globally to the entire fish via Ag-AgCl electrodes positioned ~15 cm from either side of the animal. Amplitude of this stimulus was set at 375 µV/cm rms. This same stimulus could be applied locally to small regions of the animal encompassing the receptive field of a given cell. In this case, the stimulus was applied between a ~1 mm silver ball electrode and a second electrode placed in the animal's mouth. The changes in EOD amplitude resulting from this stimulus were measured between a silver wire electrode, insulated except for the tip, placed on the skin surface within the receptive field of the cell studied and a second electrode implanted in the dorsal musculature. The resulting changes in EOD amplitude typically fell within the range of ±50-150 µV.

Stainless steel wire electrodes were used for stimulation of the dorsal and ventral molecular layer inputs to the pyramidal cell apical dendrites. An array of five 50-µm stainless steel electrodes was positioned within the DML to directly stimulate the parallel fibers (Fig. 2A1). Individual pairs of this array were selected to activate subsets of the parallel fibers providing input to a given cell. The tractus stratum fibrosum (tSF), which contains the nP efferents terminating in the VML, was stimulated with bipolar stainless steel electrodes (75-µm diam) positioned near to where the tSF exits the nP (Fig. 2B1). Single pulses (200-µs duration) or tetani (100-ms train, 200-µs pulse duration, 15-ms interpulse interval, stimulus current range 10-50 µA) were applied via WPI 360 constant current stimulus isolation units. In selected experiments, stimulating electrode positions were determined from frozen sections processed to reveal the deposition of iron via the Prussian blue reaction.


View larger version (44K):
[in this window]
[in a new window]
  		FIG. 2. A1 and B1: diagrams illustrating typical dorsal molecular layer DML and tractus stratum fibrosum (tSF) stimulation sites. Brain sections correspond to levels -8 and 1, respectively, of the A. leptorhynchus atlas (Maler et al. 1991). ALLN, anterior lateral line nerve; CCb, corpus cerebelli; cls, centrolateral segment of the ELL; cms, centromedial segment; EGa, eminentia granularis anterior; ls, lateral segment; MLF, medial longitudinal fasiculus; MOL, molecular layers; ms, medial segment; NM, nucleus medialis; PM, pacemaker nucleus; TeO, optic tectum; TS, torus semicircularis. A2: DML-evoked potential amplitudes as a function of time for different single pulse stimulation frequencies. Each point represents an average of 20 consecutive EPs. bullet , 1 Hz; black-triangle, 4 Hz; triangle , 16 Hz; open circle , 32 Hz. Inset: average of the 1st and last 20 responses to the 32-Hz stimulation (* and right-arrow, respectively) B2: tSF-evoked potential amplitudes as a function of time for the same stimulation frequencies as in A2. Recording site for the waveforms shown in the inset of A2, B2 was 443 µm below the brain surface within the DML.

Measurements of neural responses are presented as means ± SE unless indicated otherwise.

    RESULTS
Abstract
Introduction
Methods
Results
Discussion
References

Evoked potential responses

Waveforms of evoked potentials resulting from either DML or tSF (VML) stimulation closely matched those described in studies of an in vitro preparation of the ELL of Apteronotus. Potentials were initially negative within the ELL molecular layers where activation of excitatory synapses on pyramidal cell apical dendrites creates a distributed current sink. These reversed to positive below the VML where the pyramidal cell somata and basilar dendrites provide the current source (Berman et al. 1997; N. J. Berman and L. Maler, unpublished observations).

To determine the stability of pyramidal cell responses to continuous activation of molecular layer inputs, trains of 500 pulses at repetition rates ranging from 0.5 to 32 Hz were applied to either the tSF, to activate VML inputs, or to the parallel fibers of the dorsal molecular layer. Figure 2, A1 and B1, diagrammatically illustrates typical DML and tSF stimulation sites, and the plots in Fig. 2, A2 and B2, summarize the time course of changes in DML and tSF evoked potential (EP) amplitudes, respectively. Each point gives the average EP amplitude determined from blocks of 20 consecutive responses; data from five animals are averaged for the responses to 1- and 4-Hz stimulation (bullet  and black-triangle, respectively), and data from seven animals were averaged for frequencies of 16 and 32 Hz (triangle  and open circle , respectively). Stimulation of either pathway at frequencies <8 Hz evokes potentials of constant amplitude. Higher-frequency stimulation of the DML caused a rapidly facilitating EP which reached maximum amplitude within the first few presentations but then began a decay phase. With 32-Hz stimulation (open circle ) the DML-evoked EP decays to ~50% of the steady-state amplitude seen with low-frequency stimulation. An example of the effect of continuous high-frequency DML stimulation is shown in Fig. 2A2, inset. The average of the first 20 of 500 responses to 32-Hz stimulation is indicated by (*) as is that of the last 20 responses (right-arrow). The initial slope and amplitude of the latter response is significantly reduced and latency to the response peak also increased; at 32-Hz stimulation frequency, latency to peak determined from the first 20 responses averaged 6.0 ± 0.5 ms whereas that of the last 20 responses averaged 9.1 ± 0.9 ms (n = 7 animals).

Higher-frequency stimulation of the tSF (VML activation) also results in an initial rapid facilitation of the EP, however, unlike the DML, tSF responses either remain slightly facilitated (16 Hz; triangle ) or show a delayed second phase of facilitation (32 Hz, open circle ). Average responses to the first and last 20 presentations at 32 Hz are indicated in Fig. 2B2, inset, by * and right-arrow, respectively, and the latency to the response peak increased slightly from 4.9 ± 0.56 to 6.0 ± 0.55 ms (n = 7 animals).

Tetanic stimulation, 100-ms trains containing seven pulses (15-ms interpulse interval) repeated at one per second for 150 s, also was used to activate the DML and VML inputs to pyramidal cells. The records of Fig. 3, A1 and B1, show averaged responses, recorded within the pyramidal cell layer, to the first 20 presentations of the tetanus. Both DML and tSF evoked potentials show facilitation to successive stimuli within the train. Figure 3A2 shows the average of the last 20 responses to DML stimulation and, as in the case of high-frequency single pulse stimulation, responses are depressed. Responses to tetanic stimulation of the tSF did not show this strong depression (compare Fig. 3, B1 and B2). The amplitudes of the responses to the first, fourth, and seventh stimuli within the train are plotted over the 150 s of stimulation in Fig. 3, A3 and B3, for DML and tSF stimulation, respectively. With DML stimulation EP amplitude as well as the response facilitation to later stimuli within the train gradually decays to stable but reduced values within 100-125 s. However, facilitation of responses within successive presentations of the tSF tetanus remained nearly constant.


View larger version (29K):
[in this window]
[in a new window]
  		FIG. 3. A1 and A2: averages of the 1st and last 20 responses to tetanic stimulation of the DML, respectively. B1 and B2: averages of responses to the 1st and last 20 replicates of tSF stimulation. A3 and B3: average amplitudes of the 1st, 4th, and last responses evoked within each tetanus plotted over the duration of the simulation period. Each point is the average of 5 consecutive responses (pooled data from 18 animals). Recording site was within the ELL pyramidal cell layer.

Extracellular single unit responses to DML and VML stimulation

Extracellular recordings show that the responses of single ELL pyramidal cells to repeated tetanic stimulation of the DML show a progressive decay, whereas responses to the same stimulus applied to the tSF remain virtually constant. The raster diagram of Fig. 4A1 shows the responses of a nonbasilar pyramidal cell to tetanic DML stimulation; initially the cell fires in a strongly phase locked manner to the stimulus pulses but the response gradually weakens, and after ~120 s, spikes only occur on the later pulses within the train. In many cases, the initial excitatory responses to DML tetani not only decay with continuous presentation but reverse to inhibition. The histograms of Fig. 4, A2 and A4, summarize the first and last 25 1-s records, respectively, and show that the DML tetanus becomes less effective in driving this cell. The histograms of Fig. 4, A3 and A5, show the times of spike occurrence during the 10 ms after each successive stimulus within the tetanus and, as indicated by the evoked potentials, large increases in response latency occur along with the reduced responses.


View larger version (37K):
[in this window]
[in a new window]
  		FIG. 4. A1 and B1: raster displays of responses of a nonbasilar pyramidal cell to tetanic stimulation of the DML and tSF, respectively (100-ms train, 15-ms interpulse interval) at 1/s. A2 and A4, B2 and B4: poststimulus time histograms (PSTHs) summarizing responses to the 1st and last 25 replicates of the tetanus. A3 and A5, B3 and B5: PSTHs summarizing times of spike occurrence within the 10 ms after each stimulus of the tetanus for the 1st and last 25 replicates of the stimulus. A6 and B6: average spike frequencies ±SE computed from blocks of 5 consecutive responses and averaged for 26 pyramidal cells during tetanic stimulation of the DML and tSF, respectively. open circle , baseline firing frequency determined from the 300-ms periods preceding the tetanus; bullet , spike frequencies during the 100-ms tetanus.

Figure 4B1 shows the responses of the same cell to tetanic stimulation of the tSF. Activation of the ventral molecular layer inputs also caused phase locked firing of the cell, but responses remained more constant. Figure 4, B2 and B4, has histograms that summarize the first and last 25 records and show only a slight reduction in responses to the tetanus. The histograms of Fig. 4, B3 and B5, show that the latencies of the responses to the individual stimuli essentially are unchanged.

Comparisons of the effects of DML versus tSF stimulation were made for 26 pyramidal cells, and the average time course of their responses are summarized in Fig. 4, A6 and B6, respectively. No differences were seen in the behaviors of basilar versus nonbasilar pyramidal cells. Spike frequencies determined from the 300-ms periods preceding the tetanus remained constant (open circle ), however, responses to the tetanus (bullet ) differed contingent on the pathway stimulated. Responses to the DML tetanus decayed to near the level of spontaneous firing frequency, but the responses to tSF stimulation remained nearly constant. As suggested by the evoked potential experiments, extracellular single cell recordings show that responses to repeated tetanic stimulation of the DML decay but responses to tSF activation remain constant.

Responses to DML and VML stimulation show anti-Hebbian plasticity

Previous studies showed that pairing tetanic stimulation of the tSF with electrosensory stimuli applied to the receptive field of a given cell alters that cell's responses to subsequent tSF stimuli presented alone (Bastian 1996a,b). This plasticity is anti-Hebbian; pairing activation of the tSF with electrosensory stimuli that excite the cell decreases the effectiveness of subsequent tSF stimuli. Often this pairing protocol reverses the net effect of tSF stimulation from excitation to inhibition. Pairing tSF stimulation with inhibitory electrosensory stimuli increases the excitatory effects of later tSF stimuli. Experiments were done using the same pairing protocols to determine if the DML inputs to pyramidal cells are also plastic. Both pathways were studied with each cell encountered.

Tetanic stimuli were presented alone, at one per second, to either the DML or tSF (VML) until the responses to the tetanus stabilized, then responses to 30 presentations of the tetanus were recorded to determine the cell's baseline response. Examples of these initial responses are shown in the topmost segments of the rasters of Fig. 5 (DML stim. and tSF stim.). Stimulus amplitudes were selected to be near to or slightly above threshold for activating the cell. Immediately after this initial period, 60 replicates of a 100-ms stepwise increase or decrease in electric organ discharge amplitude were applied locally to the cell's receptive field synchronously with the DML or tSF tetanus (DML or tSF stim.+R. F. excit. or inhib.). After this "training period," the tetanus was again given alone at one per second for an additional 60 s.


View larger version (37K):
[in this window]
[in a new window]
  		FIG. 5. Responses of a basilar pyramidal cell to DML and tSF tetani before, during, and after pairing the tetanus with local electrosensory stimuli applied to the cells receptive field. Local stimulus increased or decreased electric organ discharge amplitude measured at the cell's receptive field by 117 µV. A1 and B1: effects of pairing inhibitory electrosensory stimuli with DML and tSF tetani, respectively. A2 and B2: effects of excitatory electrosensory stimuli paired with DML and tSF tetani, respectively.

Pairing inhibitory electrosensory stimuli with tetanic stimulation of either the DML (Fig. 5A1) or the tSF (Fig. 5B1) results in potentiation of excitatory responses to subsequent tetani presented alone. Initial spike frequencies during the DML and tSF tetani, determined from the 25 responses immediately preceding the training period, averaged 8.5 and 20 spikes/s, respectively. After the paired stimulation, average responses (25 replicates after the training) increased to 33 and 45 spikes/s, respectively. These potentiated responses gradually decayed to the pretraining levels, and this decay was generally more rapid for the DML responses.

Pairing stimulation of either pathway with excitatory electrosensory stimuli resulted in reversal of responses to subsequent tetani; pronounced pyramidal cell inhibition followed the training period (Fig. 5, A2 and B2). For the cell ofFig. 5, responses to DML tetani were reduced from 9 to 0 spikes/s and responses to the tSF decreased from 17 to 4 spikes/s. These altered responses also reverted toward their original state with continued tetanic stimulation applied alone.

Thirty-eight pyramidal cells were studied using these stimulus combinations, and their average responses are plotted over time in Fig. 6. Responses were calculated as the difference between spike frequencies determined from a 300-ms prestimulus period and the 100-ms stimulation period. Each point represents the mean response determined from five consecutive stimulus presentations averaged for the 38 cells studied. No difference in the behavior of basilar versus nonbasilar pyramidal cells was seen so data from both cell types were pooled.


View larger version (53K):
[in this window]
[in a new window]
  		FIG. 6. Summaries of the effects of local electrosensory stimuli applied to pyramidal receptive fields in conjunction with DML or tSF tetani. Responses were determined from blocks of 5 consecutive stimuli. Shaded regions indicate times of coincident pathway plus electrosensory stimulation (bullet ) or control stimulation with electrosensory stimuli alone (open circle ). A1: bullet , DML tetanus plus inhibitory stimulation of the cells' receptive fields. Pre- and posttraining responses, determined from 25 consecutive stimulus presentations before and after the paired stimulation, averaged 4.04 ± 0.72 and 21.48 ± 1.02 spikes/s, respectively (P < 0.0001, t-test, 38 cells). open circle , pre- and postcontrol responses (electrosensory stimulus alone): 6.05 ± 0.93 and 11.1 ± 1.1 spikes/s, respectively (P = 0.0006, t-test, 18 cells). A2: bullet , DML tetanus plus excitatory electrosensory stimulation, pre- and posttraining responses: 5.57 ± 0.81 and -6.38 ± 0.75 spikes/s, respectively (P < 0.0001, t-test, 37 cells). open circle , pre- and postcontrol responses: 5.34 ± 0.98 and 10.69 ± 1.04 spikes/s, respectively (P = 0.0003, t-test, 19 cells). B1: bullet , tSF tetanus plus excitatory electrosensory stimulation, pre- and posttraining responses: 8.91 ± 0.7 and 21.6 ± 0.91 spikes/s, respectively(P < 0.0001, t-test, 38 cells). open circle , pre- and postcontrol responses: 7.89 ± 0.93 and 8.71 ± 0.87 spikes/s, respectively (P = 0.52, t-test, 14 cells). B2: bullet , tSF tetanus plus inhibitory electrosensory stimulation, pre- and posttraining responses: 9.72 ± 0.74 and 0.23 ± 0.67 spikes/s, respectively(P < 0.0001, t-test, 37 cells). open circle , pre- and postcontrol responses: 5.55 ±1.03 and 5.96 ± 0.93 spikes/s, respectively (P = 0.77, t-test, 15 cells).

Responses to DML stimulation paired with inhibitory electrosensory stimuli are shown in Fig. 6A1 (bullet ). The average initial response to the tetanus alone is indicated (). Application of the inhibitory electrosensory stimulus to the cells' receptive fields (local reduction or increase in EOD amplitude for basilar and nonbasilar pyramidal cells, respectively) paired with the DML tetanus initially suppressed firing completely (shaded area) but a small recovery of responses occurred during this training period. Immediately after the training, responses to the DML tetanus were, on average, ~700% greater than the initial responses. However, these potentiated responses decayed quickly returning to baseline levels within 40 s of continued stimulation with the tetanus alone.

For a subset of these pyramidal cells (n = 18), control experiments were done in which the initial DML tetanus was followed by inhibitory electrosensory stimuli presented alone (Fig. 6A1, open circle ). The average initial responses of these cells to the tetanus is indicated (- - -). Application of the inhibitory electrosensory stimulus without the tetanus resulted in a similar suppression of responses. After the electrosensory stimulation, responses to the reapplication of the tetanus also were enhanced but significantly less so than the responses after the paired stimulation. The increased DML responses after electrosensory stimuli presented alone probably result from the pause in DML stimulation rather than the electrosensory stimulus itself. Continuous DML stimulation causes a depression of pyramidal cell responses as shown in Figs. 2-4. Removal of the tetanus during electrosensory stimulation alone allows the cell to recover from this depression, hence increased responsiveness is seen on reapplication of the DML tetanus.

The average effects of tSF tetanic stimulation paired with inhibitory electrosensory stimuli are shown in Fig. 6B1, bullet . After the paired stimulation, which suppressed pyramidal cell firing, responses to the tSF tetanus also were increased significantly. The increase in tSF responses was less than that seen for the DML, 300 versus 700%, but the potentiated responses to tSF stimulation decayed significantly more slowly. Slopes of the best-fit lines to the posttraining data of Fig. 6, A1 and B1, were -0.421 and -0.223 spikes·s-1·s-1,respectively, and these were significantly different as judged by a t-test of slopes (t = -4.21, P < 0.001). Control experiments (n = 14), in which the inhibitory electrosensory stimulus was presented without concomitant tSF tetani, resulted in no change in subsequent responses to tSF stimulation (open circle ).

Pairing either DML or tSF stimulation with excitatory electrosensory stimuli resulted in two separate effects. First, the responses to the DML or tSF tetanus did not simply sum with the responses to the excitatory electrosensory stimulus. Initial responses to DML stimulation alone averaged 5.6 spikes/s for 37 cells subsequently presented with the paired stimulation (Fig. 6A2, bullet ) and 5.3 spikes/s for a subset of 19 cells subsequently presented with the excitatory electrosensory stimulus alone (open circle ). The lines indicating these initial response levels overlap. Because the DML tetanus alone caused about a 5-spikes/s increase in firing rate and responses to the electrosensory stimulus alone averaged ~80 spikes/s (open circle , shaded region), simultaneous presentation of these stimuli should evoke increases averaging ~85 spikes/s given simple summation. Yet, when the tetanus and electrosensory stimulus were presented together, responses were only ~60 spikes/s (Fig. 6A2, bullet , shaded area). This suggests that DML stimulation also activates an inhibitory input, however, this inhibition only becomes apparent when the pyramidal cells are driven via increased electrosensory input from the periphery. This effect is equally pronounced if a pairwise analysis is done for the subset of 19 cells stimulated with both the electrosensory stimulus alone and the electrosensory stimulus plus the DML tetanus.

The second effect of coincident DML and excitatory electrosensory stimulation appears following this pairing. The initial responses to the DML tetanus were excitatory, averaging ~5 spikes/s, but after training, the tetanus caused pyramidal cell inhibition (Fig. 6A2, bullet ). This reversal of responses from excitatory to inhibitory only occurred if the tetanus was paired with the electrosensory stimulus. After the electrosensory stimulus presented without the tetanus (open circle ), DML stimulation resulted in increased excitation, and, as suggested above, this probably represents recovery from the depression due to DML tetani.

Similar results were seen when the tSF stimulation was paired with excitatory electrosensory stimuli. First, the nonlinear interaction of responses to the tetanus and electrosensory stimulation was seen (Fig. 6B2, compare open circle  and bullet , shaded area). Second, after the cessation of the paired stimulation, the responses to the tSF alone also were depressed significantly, and this depression decayed with a slower time course compared with that seen in the DML experiments. The slope of the best-fit line to the average responses after the cessation of the paired tSF plus excitatory electrosensory stimuli (0.026 spikes·s-1·s-1) is significantly less than that fit to the responses after the paired DML stimulation (0.136 spikes·s-1·s-1, t = 2.22, P < 0.05).

Pairing sensory stimulation of a given pyramidal cell's receptive field with tetanic stimulation of either the DML or tSF causes similar alterations in the subsequent responses to activation of these pathways. The changes in response show that anti-Hebbian plasticity is associated with both the DML and VML inputs to the pyramidal cells. On average, the training protocols used induce larger initial changes in the effectiveness of the DML inputs as compared with those of the tSF, however, the changes induced in the latter are more persistent.

Intracellular responses to DML and tSF stimulation

Intracellular recordings were made from ELL pyramidal cells to identify the changes in postsynaptic potentials associated with the altered effectiveness of DML and VML inputs seen during tetanus and after various pairing protocols. The excitatory or inhibitory receptive field stimulation used in the experiments of Figs. 5 and 6 was replaced with hyperpolarizing or depolarizing current injection to the pyramidal cell studied. This simplifies the experiments since receptive field locations on the animal's skin no longer need to be determined and the associated mechanical disturbances are avoided. Intracellular current injection also ensures that only the cell under study is hyperpolarized or depolarized. When receptive field stimulation is used, both the recorded cell as well as a population of neurons having similarly placed receptive fields are stimulated, allowing the possibility that local circuit phenomena contribute to the plastic effects seen.

Examples of the intracellular responses typically recorded from ELL pyramidal cells are shown in Fig. 7. A 1-s record showing responses of a basilar pyramidal cell to a single DML "test" stimulus followed by a DML tetanus is shown in Fig. 7A1. In most experiments, the test stimulus was presented along with hyperpolarizing current injection (0.1-0.3 nA), allowing the PSPs to be analyzed without contamination by action potentials. An average of 20 responses to the single DML test stimulus for the cell of A1 is shown in Fig. 7A2. Single stimuli typically evoke an initial EPSP followed by a longer duration hyperpolarization. The hyperpolarization probably reflects the increased activity of GABAergic inhibitory interneurons, including DML stellate cells and neurons of the ventral molecular layer (Maler and Mugnaini 1994; N. J. Berman and L. Maler, unpublished results). Additional examples of different pyramidal cells' responses to single stimuli show that the duration and relative amplitudes of the putative IPSPs are variable (Fig. 7, A3 and A4). This hyperpolarization often seemed to interrupt a longer-duration component of the EPSP, resulting in a second depolarization phase (Fig. 7A2, *). An average of four responses of the cell of Fig. 7A1 to the DML tetanus is shown in Fig. 7A5. An initial facilitation of successive EPSPs occurs followed by a relative depression of the last responses. This is similar to the changes in evoked potential amplitude seen with tetanic DML stimulation (Fig. 3, A1-A3).


View larger version (21K):
[in this window]
[in a new window]
  		FIG. 7. Example intracellular records from ELL pyramidal cells. A1: typical 1-s record showing the initial test stimulus occurring 50-ms after the start of a 150-ms hyperpolarization and the DML tetanus (15-ms interpulse interval) beginning 235 ms after the test stimulus. A2: average of 20 responses of this same cell to the test stimulus. A3 and A4: example postsynaptic potentials evoked by the test stimulus from 2 additional cells (averages of 20 replicates, calibration of A2 holds for A3 and A4, and B2-B4). A5: average of 4 responses of the cell of A1 to the tetanus (calibration of A5 holds for B5). B1: responses of the of the cell of A1 to tSF stimulation. B2: average of 20 responses of this same cell to the test stimulus. B3 and B4: averages of 20 responses of two additional pyramidal cells to the test stimulus. Arrows indicate the gap junction component of tSF-evoked EPSPs. B5: Average of 4 responses of the cell of A1 to the tSF tetanus.

Figure 7B1 shows responses of the same cell to tSF stimulation. The single tSF stimuli also often evoked an EPSP followed by a hyperpolarization followed by a later depolarization phase (Fig. 7B2, *), but cases in which no appreciable hyperpolarization occurred also were seen (Fig. 7B3). Similar EPSP-IPSP patterns have been observed in an in vitro preparation of the ELL (N. J. Berman and L. Maler, unpublished data) A feature unique to the tSF-evoked EPSPs is indicated by Fig. 7, B2-B4, right-arrow; this very brief initial depolarization was seen in most recordings and results from a gap junction input to pyramidal cells activated by tSF stimulation. This input has been identified morphologically (Maler et al. 1981), and in vitro studies have shown it to be insensitive to blockade by low Ca2+, high Mn2+, and it persists in the presence of both N-methyl-D-aspartate (NMDA) and nonNMDA glutamate antagonists (Berman et al. 1997). The size of this initial gap junction EPSP varied among different cells, compare Fig. 7, B2 and B4, however, this component of pyramidal cell responses was not studied further. The latencies of tSF evoked EPSPs were significantly shorter than those due to DML stimulation, (4.55 ± 0.22 vs. 8.35 ± 0.35 ms, respectively, for paired latency measures to the peak of the EPSP; n = 25 pyramidal cells, P < 0.0001, t-test). Postsynaptic responses to the tSF tetanus typically consisted of a strongly facilitating EPSP, which summated and typically exceeded spike threshold (Fig. 7B5) (Bastian 1996b; Berman et al. 1997).

Effects of single pulse stimulation paired with postsynaptic depolarization or hyperpolarization

Single DML pulses were presented continuously at 2 Hz for 150 s to gauge the stability of intracellularly recorded EPSPs. Every other stimulus (test stimulus) was paired with weak (0.1-0.3 nA) postsynaptic hyperpolarization to block action potentials and facilitate EPSP measurement. EPSP amplitudes were measured from signal averages of successive blocks of five responses to the test stimulus and expressed as the percent of mean responses to the first 30 stimuli. Average responses of 11 pyramidal cells are shown in Fig. 8A; no significant change in EPSP amplitude was seen with continuous DML stimulation at 2 Hz. Likewise, tSF-evoked EPSPs were unchanged by similar patterns of low-frequency stimulation (data not shown).


View larger version (49K):
[in this window]
[in a new window]
  		FIG. 8. Effects of postsynaptic hyper- or depolarization on responses to DML and tSF single pulse stimuli. A: average excitatory postsynaptic potential (EPSP) amplitudes evoked by continuous 2-Hz stimulation of the DML (11 cells). Each point represents the average ±SE of 5 consecutive EPSPs expressed as percent of the average of the first 30 responses. B: open circle , average EPSP amplitudes (13 cells) before, during (shaded region), and after pairing the training stimulus with 0.8-nA postsynaptic depolarization. Training stimuli occurred 65 ms after the onset of the 100-ms change in postsynaptic membrane potential. bullet , average EPSP amplitudes (11 cells) before, during, and after presentation of the postsynaptic depolarization alone. C: average EPSP amplitudes (9 cells) before, during (shaded area), and after pairing the training stimulus with postsynaptic hyperpolarization (open circle ) or presentation of the hyperpolarization alone (bullet ). D: average tSF EPSP amplitudes (5 cells) before, during (shaded area), and after pairing the training stimulus with postsynaptic depolarization and hyperpolarization (open circle  and bullet , respectively). Differences in EPSP amplitudes measured over 25 s before and after postsynaptic hyper- and depolarization averaged 2.18 ± 4.94 and 1.47 ± 4.7%, respectively, (P = 0.77 and 0.68, t-tests).

To determine the effects of pairing DML or tSF single-pulse stimulation with postsynaptic changes in membrane potential, the stimuli alternating with the test pulses (training stimuli) were paired with depolarizing or hyperpolarizing current injection. The effect of postsynaptic depolarization paired with DML stimulation is shown in Fig. 8B, open circle . Stimulation during the first 30 s was as described for Fig. 8A. After this, each training stimulus was paired with a 100-ms, 0.8-nA depolarization, which always caused high-frequency spiking of the pyramidal cell (shaded region of Fig. 8B). EPSPs evoked by the test stimulus were depressed gradually due to the paired training stimulus plus postsynaptic depolarization. A pairwise comparison of average EPSP amplitudes determined from 25 test stimuli before and immediately after the pairing gave an average EPSP reduction of 20.7 ± 3.23% (P < 0.0001, n = 13 cells, t-test). The effects of the postsynaptic depolarization applied alone was tested for 11 of these cells (Fig. 8B, bullet ). The training stimulus was stopped after the initial 30-s period and restarted after the 60 s during which the postsynaptic depolarization was applied alone. Analysis of each cell's responses preceding and after the depolarization alone showed an average increase in EPSP amplitude of 2.14 ± 3.21% (P = 0.52, n = 11 cells, t-test).

The effects of postsynaptic hyperpolarization, which always prevented the cell from producing action potentials, paired with the DML training stimuli are summarized in Fig. 8C (9 cells). No significant changes in the size of the test EPSPs were seen during the time that either the hyperpolarization plus the training stimulus or the hyperpolarization alone was applied (open circle  and bullet , respectively; shaded area). Likewise, no differential change in EPSP amplitude was seen after the paired stimulation relative to that after the hyperpolarization alone. However, compared with the pretraining values, small increases in EPSP amplitude were seen after either treatment, averaging12.9 ± 4.5 and 13.6 ± 6.8%, respectively. The increase after the paired stimulation was significant (P = 0.02, t-test), but that after the hyperpolarization alone was not (P = 0.08, t-test).

The effects of postsynaptic depolarization and hyperpolarization paired with single tSF stimuli also were studied in five pyramidal cells (Fig. 8D, open circle  and bullet , respectively). Neither treatment resulted in any significant change in the amplitude of the single tSF test pulses during or after the training periods. The test stimuli used in these and subsequent experiments was routinely paired with 0.1- to 0.3-nA postsynaptic hyperpolarization to prevent the cell from producing action potentials in response to the resulting EPSP. This treatment alone did not cause any observable change in either DML-evoked EPSPs (Fig. 8A) or in the amplitude of tSF-evoked EPSPs (data not shown).

Effects of DML and tSF tetani

Paired test stimuli were used in experiments in which tetanic stimulation of either pathway was used so that the effects on paired-pulse facilitation (PPF) could be determined. Changes in PPF often are used as an indicator of presynaptic changes in synaptic function. The twin test stimuli, 25-ms interpulse interval, were presented once per second during the 150-ms, 0.3-nA hyperpolarization used to prevent action potentials. Amplitudes of EPSPs evoked by the twin test stimuli are expressed as percent of the mean amplitude of the initial 30 responses to the first stimulus of the pair. Measurement of EPSP slopes was not attempted because in these in vivo recordings significant spontaneous synaptic activity always was present; this results in highly variable slopes. Postsynaptic potentials evoked by the DML test stimuli alone showed typical PPF averaging 170 ± 10% for 11 cells (Fig. 9A1, 0-30 s, circles). An example of EPSPs initially evoked by the paired pulses are shown in Fig. 9A2 (thin trace). After the 30 initial presentations of the test stimuli, the DML tetanus was added; its onset followed the test pulses by either 210 or 410 ms in different sets of experiments and varying this delay had no effect. The DML tetanus depressed both EPSPs evoked by test pulses (shaded region of Fig. 9A1) and examples of these recorded during the last 25 s of tetanic stimulation are shown in Fig. 9A2 (thick lines). PPF measured during the last 25 s of the tetanic stimulation fell to 144 ± 11%, and although the decrease for this data set was not statistically significant (P = 0.08, n = 11, t-test), pooling of additional data from experiments described later shows that DML tetani do cause significant reductions in PPF (see DISCUSSION). Before the tetanus, the EPSPs typically showed a steeper rise and a clear inflection, which separated the falling phase into early and late components (Fig. 9A2, arrow). The early phase of the EPSP seemed to be preferentially attenuated during the tetanus.


View larger version (43K):
[in this window]
[in a new window]
  		FIG. 9. Effects of DML and tSF tetani on pyramidal cell EPSPs. A1: EPSP amplitudes expressed as percent of initial 30 responses to the 1st test pulse before, during (shaded region), and after the addition of the DML tetanus. Circles, responses to the 1st and 2nd test stimulus, respectively. EPSP amplitudes were measured as the difference between membrane potential measured just after each stimulus artifact and at the peak of the EPSP. In some cases, the tSF-evoked EPSPs had very short latencies; in these cases, amplitude was measured as the difference between the peak and membrane potential measured just before the stimulus. Measurements were made with software which identified EPSP peaks as the largest mean of 5 consecutive voltage measurements. A2: example EPSPs evoked by the test stimuli before (thin trace) and during (thick trace) DML tetani. Records are averages of 25 replicates. Arrow, inflection point often seen separating early and late phases of DML-evoked EPSPs. B1: EPSP amplitudes before, during, and after tSF tetani. B2: EPSPs evoked by the tSF test stimuli before and during the tSF tetanus period (thin and thick lines, respectively). Calibration of A2 holds for B2.

The effect of tetanic tSF stimulation on EPSP amplitude was opposite that seen with DML stimulation. Responses to tSF twin pulse stimulation alone showed PPF essentially identical to that seen with DML stimulation (Fig. 9B1,0-30 s, PPF averaged 169 ± 12%, 9 cells). Unlike the DML, application of the tetanus further potentiated rather than depressed both the first and second responses (Fig. 9B1, shaded area). PPF also was changed significantly by tSF tetani; PPF measured during the last 25 s of the tetanus period was reduced to an average of 123 ± 4.8% (P = 0.012, t-test). Examples of the synaptic potentials evoked by the twin pulses before and during the tetanus are shown in Fig. 9B2 by the thin and thick lines, respectively. A slower hyperpolarization was also evoked in this cell by tSF stimulation, and this was increased during the tetanus as well.

The changes in EPSP amplitudes resulting from tetanic stimulation of the DML and tSF mirror the changes seen in evoked potential amplitudes and in spike responses of pyramidal cells. Repetitively activating the DML input results in a progressive depression of EPSP amplitudes, whereas responses to the tSF behave oppositely, showing strong potentiation.

Effects of tetanic DML stimulation paired with postsynaptic hyper- and depolarization

The experiment described in conjunction with Fig. 9 was modified to include a phase in which the tetanic stimulation was paired with either depolarization or hyperpolarization of the postsynaptic cell to determine if plasticity in addition to that due to the tetanus itself occurred. Figure 10, A1-A5, summarizes the effects of pairing DML tetani with postsynaptic depolarization. In the first phase of these experiments ("pre tet" region of the raster of Fig. 10A1), twin test stimuli were delivered to the DML alone, at one per second, and the baseline EPSP amplitudes were measured. After this, the DML tetanus was introduced, as described in the previous section, and this continued for an additional 120 s ("tet1" region, Fig. 10A1). In the third phase of this experiment, 0.8-nA depolarizing current injection was delivered to the pyramidal cells in conjunction with the DML tetanus for 60 s ("tet + dep" region, Fig. 10A1). After this training paradigm, the DML tetanus again was delivered alone (Fig. 10A1, "tet2" region).


View larger version (43K):
[in this window]
[in a new window]
  		FIG. 10. A1: raster display summarizing pyramidal cell spike responses to tetanic stimulation of the DML before (tet1), during (tet + dep.), and after (tet2) addition of addition of 0.8-nA depolarizing current injection. A, 2 and 3: individual records of this cell's responses to the DML twin test stimuli immediately preceding and following the paired stimulation. A4: averages of 25 responses to the test stimuli before (thin) and after (thick trace) the paired stimuli. A5: averages of 25 responses to the DML tetanus before and after the paired stimulation (thin and thick traces, respectively). Before averaging, spikes were removed from the original records. This was done with an algorithm that identified the time of occurrence and duration of the action potentials in digitized records and replaced these regions of the records with the average of membrane potential values immediately preceding and following the spikes. This technique provides a less cluttered view of slow potential changes but it also reduces the apparent amplitude of individual EPSPs within the tetanus. This procedure was not applied to any records of responses to twin test pulses. B, 1-5: responses of the same cell before, during, and after DML tetanus paired with postsynaptic hyperpolarization.

Comparison of the tet1 and tet2 regions of the raster show that the effects of the DML tetanus reversed from excitation to inhibition after pairing with postsynaptic depolarization. Examples of single 1-s records recorded immediately before and after the application of the paired stimuli (Fig. 10, A2 and A3, respectively) show responses to the twin test stimuli and to the DML tetanus. Comparison of averages of 25 test EPSPs recorded before and after the pairing show that this treatment strongly depressed EPSP amplitudes (Fig. 10A4, thin and thick lines, respectively). Depression of EPSP amplitude alone, however, is not sufficient to account for the reversal of the spike responses to the tetanus from excitation to inhibition. Averages of 25 responses to the DML tetanus before and after the paired stimulation (Fig. 10A5, thin and thick traces, respectively) show that a large hyperpolarization develops after the paired stimulation and this hyperpolarization prevents the DML-evoked EPSPs from reaching spike threshold. Action potentials were removed from the records averaged in Fig. 10A5, and this procedure results in an underestimation of the peaks of individual EPSPs but doesn't significantly affect the slower membrane potential changes.

Figure 11A (filled circles, lightly shaded area) summarizes the average time course of changes in the first test EPSP's amplitude averaged for 11 cells, and as shown in Fig. 9A1, the DML tetanus depresses these responses. Adding the postsynaptic depolarization (heavy shading) depresses EPSP amplitudes well below the level resulting from the tetanic stimulation alone. This second phase of EPSP depression is reversed after removal of the postsynaptic depolarization even though the DML tetanus was continued.


View larger version (47K):
[in this window]
[in a new window]
  		FIG. 11. A: changes in DML-evoked EPSP amplitudes during tetanic stimulation (circles and light shading), during tetani paired with postsynaptic depolarization (filed symbols, heavy shading, 12 replicates, 11 cells), and during tetani paired with postsynaptic hyperpolarization (open symbols, heavy shading, 13 replicates, 11 cells). For each cell, EPSP amplitudes were measured from means of successive blocks of 5 responses. Each point represents the average ± SE of responses expressed as percent of the mean response to the first 30 replicates of the test stimulus given alone. B: histograms summarizing average EPSP amplitudes evoked by the first (EPSP1) and second (EPSP2) test stimuli determined from 25 stimulation trials before the tetanus (open bars) and immediately before and after the paired tetanus plus postsynaptic depolarization (grey and black bars, respectively). Reductions in EPSP amplitudes due to the tetanus alone averaged 1.3 ± 0.26 and 2.4 ± 0.4 mV for the 1st and 2nd test EPSPs, respectively. Both reductions are significant (P <=  0.0004, t-tests). Reductions in EPSP amplitudes due to addition of the postsynaptic depolarization averaged 0.73 ± 0.24 and 0.81 ± 0.22 mV, respectively, and these additional reductions also are significant (P = 0.009 and 0.003, t-tests). Hatched bars show average spike responses to the DML tetanus before (pre, open bars) and after (post, black bars) the paired stimulation. C: average EPSP amplitudes evoked by the 1st and 2nd test stimuli before the onset of the tetanus, before the tetanus plus hyperpolarization and after this treatment (open, hatched, and black bars, respectively). Reductions in the 1st and 2nd test EPSP amplitudes due to the tetanus averaged 1.7 ± 0.49 and 2.5 ± 0.32 mV, respectively (P = 0.005 and <0.0001, t-tests). After DML stimulation paired with postsynaptic hyperpolarization, EPSPs increased by 0.54 ± 0.12 and 0.43 ± 0.11 mV, respectively (P = 0.0007 and <0.0001). Hatched bars show average spikes responses before and after the paired stimulation.

The histogram in Fig. 11B shows the average amplitudes of first and second test EPSPs (EPSP1 and -2) before the onset of the tetanus (open bars), immediately preceding the tetanus plus depolarization (grey bars) and after this paired stimulation (black bars) for the 11 cells studied. Analysis of the average changes in EPSP amplitudes showed that the depression resulting from the tetanus alone as well as that due to the tetanus plus the depolarization are highly significant (see Fig. 11, legend). PPF, determined from the data of Fig. 11B, had an initial value of 150 ± 6.8%, this was reduced by 19% to 131 ± 13.4% and, as with the data of Fig. 9, the reduction of PPF was marginally significant(P = 0.06, t-test). Although the combined tetanic stimulation plus depolarization resulted in additional depression of EPSP amplitudes, PPF showed no further change, averaging 139 ± 14.4%, after the paired stimulation.

The hatched bars of Fig. 11B show average spike responses to the DML tetanus for these cells determined from the 25 responses immediately before (lightly hatched bar) and after (black hatched bar) the paired stimulation. The responses to DML tetani consistently reversed from excitation to inhibition after the pairing with postsynaptic depolarization. This inhibition of spike responses results from the development of DML evoked hyperpolarization after pairing as shown in Fig. 10A5.

The effects of paired DML stimulation and postsynaptic hyperpolarization are summarized in Fig. 10B for this same pyramidal cell. Comparison of the tet1 and tet2 regions of the raster display (Fig. 10B1) shows that after the postsynaptic hyperpolarization plus DML tetanus, spike responses are increased. After pairing, the cell responds in a phase-locked manner to nearly every pulse within the tetanus (compare Fig. 10, B2 and B3). The averages of 25 test EPSPs before and after paired stimulation shows that this treatment nearly doubled EPSP amplitudes (Fig. 10B4, thin lines and thick lines, respectively), and the averages of responses to the DML tetanus also show increased depolarization after the paired stimulation (Fig. 10B5, thin and thick lines, respectively). Spikes were removed from these records before averaging; this causes underestimation EPSP amplitudes, hence true differences in EPSPs amplitudes during the tetanus maybe underestimated.

The average EPSP amplitudes for 13 experiments of this type are plotted as open symbols in Fig. 11A. Pairing the DML tetanus with postsynaptic hyperpolarization (heavy shading) largely reversed the EPSP depression due to tetanic stimulation alone (light shading). The histogram of Fig. 11C shows average EPSP amplitudes evoked by both the first and second test pulses before the tetanus (open bars) and before and after the paired tetanus plus hyperpolarization (gray and black bars, respectively). Analysis of the changes in test EPSP amplitudes shows that the addition of postsynaptic hyperpolarization to the tetanus results in a significant reversal of the depression caused by the tetanus alone (Fig. 11, legend). As seen in the earlier data, the DML tetanus alone reduced PPF ~20%, from 149 ± 6.9% to 130 ± 8.3% (P = 0.06, t-test). The hatched bars of Fig. 11C show that, on average, pairing the DML tetanus with postsynaptichyperpolarization increased spike responses to subsequenttetani ~10-fold.

These data show that the amplitude of DML-evoked postsynaptic potentials are modulated by coincident presynaptic activity and changes in the postsynaptic cell's membrane potential. Presynaptic activity paired with postsynaptic depolarization results in depression of EPSP amplitude as well as potentiation of probable DML-evoked IPSPs. Coincident postsynaptic hyperpolarization has the opposite effect; EPSPs clearly are potentiated, and it is also possible that this treatment depresses IPSP amplitudes.

Effects of tetanic tSF stimulation paired with postsynaptic hyper- and depolarization

Pairing tSF tetani with postsynaptic depolarization changed responses from excitation to inhibition as was seen with DML stimulation (compare tet1 and tet2 regions of Fig. 12A1 and Fig. 13B, pre- and postspike, respectively). For the cell of Figs. 10 and 12, the change in spike response due to the paired depolarization was associated with a depression of the second test EPSP, but not the first (Fig. 12A4, compare thick and thin trace), and with a reduction in the summated depolarization seen during the tSF tetanus (Fig. 12A5, thick vs. thin trace). Figure 13A (filled symbols) shows average changes in the amplitude of the first test EPSP for 11 experiments (10 cells) of this type. As shown earlier, tetanic simulation of the tSF alone results in significant potentiation of the EPSPs evoked by the test stimuli (Fig. 13A, lightly shaded area; Fig. 13, B and C, clear and lightly shaded areas). The average effect of pairing the tSF tetanus with postsynaptic depolarization is a partial reversal of the potentiation due to the tetanus alone as shown in Fig. 13A (filled symbols and heavy shading) and Fig. 13B (black bars). Although this averaged data shows a significant decrease in the amplitude of the test EPSPs, there was some variability in the responses of individual cells. Of the 10 neurons studied, 5 showed little or no depression of the first test EPSP when averages of 25 EPSPs immediately before and after the paired stimulation were compared, however, the second test EPSP always was depressed. The remaining five cells showed larger reductions of both test EPSPs. In an earlier study of the tSF plasticity, single test stimuli were used, and no significant depression after paired tSF tetani and postsynaptic depolarization was seen (Bastian 1996a). Further studies are needed to determine whether differences in tSF stimulation sites, individual variation among pyramidal cells, or some other unknown factor accounts for this variability.


View larger version (48K):
[in this window]
[in a new window]
  		FIG. 12. Summary of responses of the pyramidal cell of Fig. 10 to tSF stimulation paired with postsynaptic depolarization (A1-A5) and paired with postsynaptic hyperpolarization (B1-B5). Descriptions of Fig. 10 apply to these data; spikes were removed from the records averaged in A5 and B5.


View larger version (43K):
[in this window]
[in a new window]
  		FIG. 13. A: changes in tSF-evoked EPSP amplitudes during tetanic stimulation (circles and light shading), during tetani paired with postsynaptic depolarization (filled symbols, heavy shading, 12 replicates, 11 cells), and during tetani paired with postsynaptic hyperpolarization (open symbols, heavy shading, 16 replicates, 12 cells). B: average amplitudes of test EPSPs (EPSP1 and EPSP2) before the tetanus (open bars), during the tetanus but before postsynaptic depolarization (grey bars), and during tetanus but after the postsynaptic depolarization (black bars). Increases in EPSP amplitudes due to the tetanus averaged 1.7 ± 0.2 and 0.9 ± 0.31 mV for EPSP1 and -2, respectively (P = 0.0001 and 0.013, t-test). Reductions due to the paired depolarization averaged 0.61 ± 0.16 and 0.69 ± 0.14 mV (P = 0.003 and 0.0007, t-test). Before the tetanus PPF averaged 143 ± 13.7%, It was reduced to 114 ± 5.5% during the tetanus (P = 0.043 t-test) and after the paired depolarization PPF averaged 113 ± 6.4%. The addition of the postsynaptic depolarization had no significant effect on PPF (P = 0.99,t-test). Hatched bars show changes in spike responses before (pre) and after (post) the paired stimulation. C: average EPSP amplitudes before tetanus and before and after the tetanus paired with postsynaptic hyperpolarization (open, grey, and black bars, respectively). Increases in the 1st and 2nd test EPSP amplitudes averaged 1.33 ± 0.19 and 0.41 ± 0.15 mv, respectively (P < 0.0001 and = 0.017, t-test). Increases in EPSP amplitudes after paired tSF tetanus plus postsynaptic hyperpolarization averaged 0.41 ± 0.09 and 0.33 ± 0.07 (P = 0.0002 and <0.0001, t-test). Before the tetanus, PPF averaged 182 ± 12.1%. It was reduced to 99.9 ± 4.9% during the tetanus (P < 0.0001, t-test), and after the paired hyperpolarization, PPF averaged 98.5 ± 4.9% (P = 0.72, t-test). Hatched bars show average changes in spike response before and after the paired stimulation.

Despite the variability seen in the effects of paired tSF tetani and postsynaptic depolarization on the first test EPSP, the effect of this treatment on pyramidal cell spike responses is clear; these reversed from excitation to inhibition (Fig. 13B, hatched bars), and in many cases, this reversal in spike responses was associated with the appearance of a tSF-evoked hyperpolarization as well as reduced EPSP amplitudes during the tetanus.

After tSF tetani paired with