Optical recording of trisynaptic pathway in rat hippocampal slices with a voltage-sensitive dye

Pergamon

PII:

Neuroscience Vol. 81, No. 1, pp. 1–8, 1997 Copyright ? 1997 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/97 $17.00+0.00 S0306-4522(97)00161-9

OPTICAL RECORDING OF TRISYNAPTIC PATHWAY IN RAT HIPPOCAMPAL SLICES WITH A VOLTAGE-SENSITIVE DYE Y. NAKAGAMI, H. SAITO and N. MATSUKI* Department of Chemical Pharmacology, Faculty of Pharmaceutical Sciences, University of Tokyo, Tokyo, Japan Abstract––Changes in membrane potentials were recorded from rat hippocampal slices with a voltagesensitive dye using a real-time optical recording system, which had high spatial resolution of 128#128 points with a high time resolution of 0.6 ms. Serial excitatory propagation was recorded in the dentate gyrus, CA3 and CA1 after stimulation of the perforant pathway, and the optical signals were clearly divided into two components in the dentate gyrus adjacent to the stimulus site. The slow component was suppressed in Ca2+-free solution, but the fast component in the molecular layer was not affected. However, the application of 1 µM tetrodotoxin fully abolished both components. These results suggest that the fast and slow components mainly reflect Na+-dependent action potentials and excitatory postsynaptic potentials, respectively. The excitatory response duration in the stratum radiatum of CA3 was significantly longer than that in other hippocampal areas. The long-lasting excitation in CA3 is probably related to the CA3 associational projections, because direct stimulation of CA3 pyramidal cell layer also produced similar results. The long-lasting dendritic excitation is probably important to integrate synaptic transmission and may be related to epileptogenesis. When long-term potentiation was induced by a tetanic stimulation (100 Hz for 1 s), the onset latency in the stratum radiatum of CA1 was reduced to as much as 65%, suggesting an increase of excitatory propagation. The analysis of the spatial–temporal optical signals contributes to understanding information processes in the hippocampus, related to learning and memory including long-term potentiation. ? 1997 IBRO. Published by Elsevier Science Ltd. Key words: hippocampus, optical recording, voltage-sensitive dye, long-term potentiation.

The hippocampus is a subcortical structure that has been associated with learning and memory. To date, analysis of the main neuronal networks of the hippocampus has emphasized its intrinsic trisynaptic pathway. The hippocampus consists of three major subdivisions, namely the dentate gyrus (DG), the CA3 region and the CA1 region. The perforant input induces serial excitatory propagation in DG, CA3 and CA1. This so-called trisynaptic pathway has been considered to be the fundamental network of the hippocampus and is thought to be involved in neuronal information processing.2 However, only a few electrophysiological studies concerning this trisynaptic pathway have been reported,28 presumably because of difficulties of simultaneous recording using multiple conventional electrodes and availability of information only from relatively large cell bodies and thick dendrites. Optical recording with voltage-sensitive dyes is one strategy for overcoming this problem.5,7,14,22 This system can monitor neuronal activities at multiple

sites simultaneously and has spatial and temporal resolution. The real-time recording system which we employed is feasible to record at 128#128 points with a high time resolution of 0.6 ms.17 The system enabled us to investigate electrical activity in the piriform cortex,23,24 visual cortex26 and neurohypophysis.20 Recently, Iijima et al.11 demonstrated signalling in the entorhinal neuronal circuit into the hippocampus using a similar system and we also reported the rat entorhino-hippocampal system in organotypic culture.18 We applied this optical image recording system to rat hippocampal slices to investigate the trisynaptic pathway and long-term potentiation (LTP), which is a form of synaptic plasticity regarded as a model of learning and memory.19 EXPERIMENTAL PROCEDURES

Slice preparation and electrophysiological recording Male Wistar rats (SLC, Japan) weighing 130–170 g were rapidly decapitated, and transverse slices of hippocampus, 400 µm thick, were prepared using a microslicer (DTK1500, Dosaka E. M., Japan). Slices were maintained at 33)C for at least 1.5 h prior to recording in artificial cerebrospinal fluid (ACSF) containing (in mM) NaCl 127, KCl 1.6, KH2PO4 1.24, MgSO4 1.3, CaCl2 2.4, NaHCO3 26 and glucose 10, which was continuously bubbled with a mixture

*To whom correspondence should be addressed. Abbreviations: ACSF, artificial cerebrospinal fluid; CNQX, 6-cyano-7-nitroquinoxaline; DG, dentate gyrus; LTP, long-term potentiation; NMDA, N-methyl--aspartate; TTX, tetrodotoxin. 1

2

Y. Nakagami et al.

of 95% O2 and 5% CO2. Then, a slice was transferred to a recording chamber and perfused at a rate of 2.0 ml/min with ACSF at room temperature (25–27)C). Conventional electrophysiological techniques for extracellular recordings were employed to check the responses of the slice and to adjust the stimulus strength. A glass electrode filled with 0.9% NaCl and a pair of stimulating electrodes (interpolar distance about 150 µm) were placed on the granule cell layer and the molecular layer of DG, respectively, for simultaneous electrical recording. A rectangular pulse (0.2– 0.7 mA) of 100 µs duration was delivered every 15 s with the strength that gave 70–80% of maximum spike amplitude. A train of 100 pulses at 100 Hz was used as tetanic stimulation. Optical recording A slice was stained for 3.5 min in 2.0 mg/ml RH155 solution1,3,4,15,18,27 dissolved in ACSF and then washed in normal ACSF for at least 15 min. The recording chamber was mounted on an inverted microscope (IMT-2, Olympus, Japan) that was equipped with a 12 V/100 W tungsten/ halogen lamp, heat absorption filter, interference filter, a mechanical shutter and objective lenses. The optical recording system (HR Deltaron 1700, Fuji Photo Film, Japan) consisted of a 128#128 photodiode array and a processing unit.17 Transmitted light with a wavelength of 700&20 nm was projected. The duration of light exposure was limited (1 s) by a mechanical shutter to avoid photodynamic damage and dye bleaching. Using a #4 objective lens, each photodiode array element received light from 17.5#17.5 µm2 area of hippocampal tissue, and the whole area of measurement was equal to 2.24#2.24 mm2. Before stimulation, a 128 frames averaged image of background light signal was stored in reference memory. These stored reference data were continuously subtracted from real-time images during the measurement, and transferred to the unit sequentially at a frame rate of 0.6 ms. One trial consisted of 512 real-time images, and 16 trial images were averaged to improve the signal-to-noise ratio. All images were recorded on condition of camera gain 1 and amp gain 1000 (HR Deltaron 1700 software Ver. 1.22). The peak amplitude, the latency (the time from the stimulus artifact to the onset), the duration (the time from the onset to the recovery point) and the slope (linking the onset and the peak point) are measured by extrapolating a baseline. All results are expressed as means&S.E.M. Materials RH155 was purchased as NK-3041 from Nippon Kankoh-Shikiso Kenkyusho (Japan), tetrodotoxin (TTX) was from Research Biochemical Inc., and 6-cyano-7nitroquinoxaline (CNQX) was from Sigma Chemical Co. RESULTS

Excitatory propagation of trisynaptic pathway Figure 1 indicates the approximate outline of pyramidal cells and granule cells of the hippocampal slice and the dotted square corresponds to the actual measurement area in this study. As shown in Fig. 2, stimulation of the perforant pathway resulted in the sequential propagation of the optical signals in DG, CA3 and CA1 fields. We believe that these optical signals primarily reflect changes in membrane potentials of the neurons and processes (see Discussion). The onset latency to sharp deflection of the optical signals was progressively longer in this order:

Fig. 1. Schematic diagram of rat hippocampal slice. The dotted square is equal to the actual measurement area in this study. pp, perforant pathway; mf, mossy fibre; sc, Schaffer collaterals; stim, stimulating electrodes; rec, recording electrode.

5.5&1.0 ms for granule cell layer, 13.5&1.3 ms for CA3 pyramidal cell layer, 19.2&1.2 ms for CA1 pyramidal cell layer (n=6). The latency measured optically corresponded well to those recorded electrically (data not shown). The feasibility of making simultaneous recordings from multipoints (128#128) and the imaging of signal propagation are great advantages of the optical recording technique. Near the cathodal stimulation electrode (i.e. depolarized site), a long-lasting depolarized signal was observed, but was absent more than 100 µm away from the site (the location is indicated as an asterisk in Fig. 2). The duration of this signal sometimes exceeded 300 ms. The signal is probably an artifact due to the local depolarization or alteration of voltage-sensitive dye properties, because it tended to be graded with a decrease in stimulus intensity.8 Characterization of optical signals The time course of the optical signals in DG adjacent to the stimulus site after perforant pathway stimulation is shown in Fig. 3 and each trace was obtained from every two photodiodes by averaging 16 sweeps. The dotted squares are equal to the area each photodiode covers. At points adjacent to the stimulus site, the optical signals were definitely divided into two components: fast and slow (left column traces in Fig. 3). The fast component had short latency (2–5 ms) and its duration was about 2–8 ms. In contrast to the fast component, the slow component had 5–8 ms latency and its duration was in the range of 40–150 ms. Away from the stimulus site, the difference between the two components gradually disappeared, presumably because synchronization of discharges decreased. As shown in the left column traces of Fig. 3, the slow component in the inner third of the molecular

Optical recording of hippocampal trisynaptic pathway

3

Fig. 2. Optical signal propagation in the trisynaptic pathway from hippocampal preparation. The fractional absorbance change was greyscale-coded, as shown by the scale in the bottom right corner. Dashed lines in 0 ms image indicate the approximate outline of pyramidal and granule cell layers in hippocampal slices, and an asterisk indicates the stimulus site.

layer of DG was stronger than in other regions after stimulation of the perforant pathway. The Ca2+ dependency of the components was investigated to characterize the optical signals. In Ca2+-free solution (replaced with Mg2+), the fast component was slightly reduced in the polymorphic and granule cell layers, but the slow component was strongly depressed to 0–10% of the basal level, as shown in the middle column traces of Fig. 3. The extracellular recording was also continuously monitored and showed no population spike in Ca2+-free solution 10 min after replacing the solution. Although this depression was reversible and the amplitude of the fast and slow components increased again to the basal level, a few synchronized oscillatory signals were often observed during recovery (data not shown). Similar results were observed in the CA3 and CA1 regions when the mossy fibre and the Schaffer collaterals were stimulated, respectively. The fast component seemed to reflect mainly Na+dependent action potentials, because it was not affected significantly in the molecular layer in Ca2+free solution. To test this idea, the slice was perfused with an Na+ channel blocker, TTX. Both the fast and slow components were completely blocked by 1 µM TTX as shown in right column traces of Fig. 3. This observation agrees with the previous reports,8,15 and suggests that the fast component represents the occurrence of action potentials in the stimulated axons.

Long-lasting excitation in CA3 Figure 4 indicates the time course of the optical signals in CA1, CA2 and CA3 after perforant pathway stimulation, and the image corresponds to the optical pattern 60.0 ms after the stimulation and indicates locations from where the traces were obtained. Although the trisynaptic circuit is recognized as the fundamental circuit of the hippocampus, the CA2 field also showed a weak excitatory response. The amplitudes of the optical signals in pyramidal cell layer were weaker than those in stratum radiatum. This may reflect a lower ratio of surface membrane to cell volume in the area comprised mainly of somata, which would cause the quantity of dye per pixel to be smaller. Because pyramidal cells in CA1 receive the Schaffer collateral input from pyramidal cells in CA3, it was expected that CA1 region would be more difficult to activate synchronously compared to CA3 region, causing the signal in CA1 to broaden and be weaker than the signal in CA3. However, the slope and duration of the optical signals in CA3 ((1.00&0.05)#10-3%/ms and 136.9&3.8 ms, n=6) were more gradual and long-lasting than those in CA1 ((1.79&0.11)# 10-3%/ms and 92.4&9.0 ms, n=6). What makes the difference between pyramidal cells of the two regions with regard to the signal waveform? Anatomical studies suggest the possibility that associational connections may be a cause of the genesis of the

4

Y. Nakagami et al.

Fig. 3. Time-course of optical signals in DG adjacent to the stimulus site. Nine traces from each pixel were shown every two photodiodes averaging 16 sweeps. The left column traces were obtained in normal solution. The middle column traces represent the optical signals in Ca2+-free solution from the same points. Please note a marked reduction of the slow component. The right column traces were obtained after application of 1 µM TTX. The diagram of a granule cell in DG shows the relative positions from which the traces were obtained. The dotted squares in this scheme are equal to the monitoring area each photodiode covers and the location of the stimulus site was placed about 150 µm away from this area. These signals represent the average of 16 sweeps obtained with an interval of 15 s. Stimulus delivery is indicated by arrowheads.

Fig. 4. Time-course of optical signals in CA1, CA2 and CA3. The time image in the centre was obtained 60.0 ms after perforant pathway stimulation. Note the longer duration and more gradual onset latency of CA3 signal than those of CA1. Stimulus delivery is indicated by arrowheads.

difference. Multiple excitatory postsynaptic potentials, including action potentials, could arise in CA3 more readily than in CA1, because the associational connections from CA3 to CA3 are organized in a systematic fashion.

Therefore, we examined whether the associational projections would be activated more by intense monosynaptic stimulation. In Table 1, the duration following the perforant pathway stimulation indicates the duration of the excitatory response in the

Optical recording of hippocampal trisynaptic pathway Table 1. Duration of the excitatory response in the granule cell layer of dentate gyrus, the stratum radiatum of CA3 and of CA1

CA1 CA3 DG

Duration following perforant pathway stimulation (ms)

Duration following monosynaptic stimulation (ms)

105.7&4.9 136.9&3.8** 92.4&9.0

104.1&7.8 182.0&7.7## 109.3&6.7

The duration of the excitatory response corresponds to the duration of each field after perforant pathway stimulation with 70–80% of maximum intensity or monosynaptic stimulation with maximum intensity, respectively. Results are expressed as means&S.E.M., n=6. **P<0.01 vs duration after perforant pathway stimulation in DG (Dunnett’s multiple range test), ##P<0.01 vs duration after perforant pathway stimulation in CA3 (Student’s t-test).

granule cell layer of DG, the stratum radiatum of CA3 and of CA1 after the perforant pathway was stimulated with 70–80% of maximum intensity. The duration after monosynaptic stimulation indicates the duration in each field after the perforant pathway, the mossy fibre or the Schaffer collaterals were stimulated with maximum intensity, respectively. Only the duration in CA3 showed significant change following monosynaptic stimulation and it exceeded 150 ms. In DG, the duration seemed to become longer following monosynaptic stimulation, but the difference did not reach statistical significance. If the associational projections were connected with the long-lasting excitation within the CA3 field, the direct stimulation of CA3 pyramidal cell layer would induce similar optical signals. As expected, the direct stimulation of CA3 pyramidal cell layer also evoked similar long-lasting signals in the stratum radiatum of CA3 (Fig. 5A, n=6). We then applied a non-N-methyl--aspartate (NMDA) glutamate receptor antagonist, CNQX, and the optical signal was depressed about 10 min after the application of 10 µM CNQX. The result suggests that the response in CA3 was not due to the direct stimulation of CA3 pyramidal neurons, but reflected a postsynaptic response. To rule out the possibility that the recurrent input of CA1 to CA3 region affected the excitatory duration of CA3 after the action potentials in the Schaffer collaterals propagated to CA1, the effect of cutting the Schaffer collaterals was investigated. The CA1 response was fully diminished by cutting the Schaffer collaterals, but CA3 response showed no apparent change (Fig. 5B). Similar results were obtained in three independent experiments. Effects of long-term potentiation Finally, the effect of LTP induction on the optical signals was investigated. This induction altered the

5

shape of the optical signal waveforms. The signal peak amplitude and latency changed in parallel with the alteration of the electrical recording in DG, although they did not always increase similarly in CA3 and CA1. Table 2 summarizes the analysis of the optical signal waveforms before and after induction of LTP and the onset latency in CA1 was reduced to about 65%. The time image in Fig. 6 was obtained 54.0 ms after perforant pathway stimulation. The expansion of the response area after LTP was consistently observed, but we do not yet have a quantitative measure of the increase. DISCUSSION

Characterization of optical signals At points adjacent to the stimulus site, the optical signals were clearly divided into two components. It is reported that the fast and slow signals represent action potentials and excitatory postsynaptic potentials, respectively, in the rat hippocampal slice.8 We also demonstrated the Ca2+ dependency of the slow component and TTX sensitivity of the fast component. But our results indicate that the fast component in the polymorphic and granule cell layers of DG would consist not only of action potential but also Ca2+ component of the spikes, because the fast component in those layers was depressed slightly in Ca2+-free solution. In addition, both components were recovered to the basal level after washout, therefore the slight decrease was not due to the bleach of a dye. The size of the slow component was maximal in the inner third of the molecular layer. The entorhinal afferents mainly innervate the outer two-thirds of the molecular layer, and the commissural and associational excitatory inputs project to more proximal somata in the inner third of the molecular layer.9,10 It is likely that the slow component in the molecular layer is related to excitatory associational circuit. However, the large size of the optical signals may be due to a large amount of dye and the large surface area of the dendrites in the region. Because the optical signals represent summed activity throughout the thickness of the tissue in the area each photodiode covers and are not only dependent on the magnitude of the membrane potential, but also the synchronicity of the activity in the area, the separation between the two signals is less clear away from the stimulus site. Therefore, the signals in DG, CA3 and CA1 appear to have only one component, as shown in Fig. 6. The involvement of glial depolarization in the slow component has been suggested. In skate cerebellar slices, RH155 showed fast and slow components in the same way, and RH482, a closely related pyrazo-oxonol dye, exhibited only fast component.15 The experimental data suggested that the two dyes have similar affinities for the parallel fibre membrane, and that RH155 has more affinity for one or more

6

Y. Nakagami et al.

Fig. 5. The long-lasting excitation in CA3. (A) Effects of a non-NMDA glutamate receptor antagonist, CNQX, on the optical signals in the stratum radiatum of CA3 caused by the direct stimulation of CA3 pyramidal cell layer. Optical signals were depressed about 10 min after the application of 10 µM CNQX. (B) Effects of cutting the Schaffer collaterals (sc cut) on the optical signals in the stratum radiatum of CA1 and CA3. Note that CA1 signal was fully diminished by cutting the Schaffer collaterals, but CA3 signal showed no apparent change. Stimulus delivery is indicated by arrowheads. Table 2. Change of optical signal waveforms before and after induction of long-term potentiation Control

CA1 CA3 DG

LTP

amplitude (%)

latency (ms)

amplitude (%)

latency (ms)

0.049&0.005 0.045&0.006 0.055&0.005

19.4&1.3 12.0&0.8 5.4&0.7

0.063&0.008* 0.053&0.008* 0.065&0.004**

12.8&1.3* 9.0&1.0* 5.0&0.7

Results are expressed as means&S.E.M., n=5. *P<0.05, **P<0.01 vs each control amplitude or latency (Student’s t-test).

Fig. 6. Time-course of optical signals from a point adjacent to the stimulus site, the molecular layer of DG, stratum radiatum of CA3 and CA1, before and after tetanic stimulation. The time image in the centre was obtained 54.0 ms after tetanic stimulation of the perforant pathway. The onset latency of CA1 became much shorter after tetanic stimulation. The optical signals before and after induction of LTP are superimposed in the right column. Stimulus delivery is indicated by arrowheads. The facilitation continued for more than 1 h.

other types of cells such as glial cell membrane in skate. However, Cinelli et al.4 reported that the enhancement of the slow component arose from

mitral/tufted process by long-lasting regenerative Ca2+ current using a salamander olfactory bulb. They proposed glial depolarization, because

Optical recording of hippocampal trisynaptic pathway

application of K+ channel blocker resulted in broadening of both components. The appearance of a delayed deporalization of CA1 neurons and other adjacent cells was also suggested by the application of 4-aminopyridine in rat hippocampal slice.3 The optical signals recorded in the present study would mainly reflect changes in membrane potential of the neurons for the following reasons: (i) no change in optical signals was detected in unstained slices, (ii) optical and electrical recordings showed identical latency and excitation threshold, (iii) similar changes of optical signals were already reported using the same type of dye,1,3,4,11,15,27 (iv) elimination of extracellular Ca2+ and addition of TTX affected the signals. However, our present data cannot rule out possible contribution of glial depolarization to the optical signals, therefore further studies that employ pharmacological analysis and different approaches such as simultaneous measurement of Ca2+ influx and optical recording of membrane potential are required. Long-lasting excitation in CA3 In the present study, we demonstrated a trisynaptic propagation of excitation in the hippocampus using optical recording suggested by the increase in onset latency with distance from the stimulating electrodes. The slope and duration of the optical signals became steeper and shorter after the propagation from CA3 to CA1. In CA3 region the excitatory signal was long-lasting and it became even longer after monosynaptic stimulation. Anatomical studies showing dense CA312,16 and sparse CA125 associational projections suggest a possible relationship between the long-lasting signals and associational connections. We demonstrated that direct stimulation of CA3 pyramidal cell layer also induced long-lasting signals that may reflect synaptic responses. These results probably suggest the connection of CA3 associational projection in long-lasting excitation. Since interictal-like activity was found in the CA2–CA3 region,6,13 the prolonged excitation in CA3 may

7

indicate that CA3 region is more likely to initiate epileptiform activity than other regions. In addition, we used only the medial part of the temporal side of hippocampal slices in these experiments. Therefore, further studies in other regions of hippocampus are necessary. Effects of long-term potentiation LTP has not been analysed as a spread of spatial propagation of excitation using optical recording. Saggau et al.21 reported an increase of optical signal after the induction of LTP on Schaffer collateral– CA1 synapses. We demonstrated trisynaptic spread of excitation after LTP and the decrease of onset latency of the optical signals, especially in CA1, indicating an augmentation of overall propagation. Since a change of conduction velocity in axons is not likely, the decrease in the latency probably indicates that synaptic transmission is speeded up after the induction of LTP. Iijima et al.11 also reported the entry of neuronal activity from the entorhinal cortex to the hippocampus in a frequency-dependent way using the same system.

CONCLUSIONS

The present results demonstrate (i) signal propagation of the hippocampal trisynaptic pathway by optical recording, (ii) long-lasting excitation in CA3 field, and (iii) dramatic decrease of the onset latency in CA1 by trisynaptic LTP induction. Some evidence was presented suggesting that associational projections were related to CA3 long-lasting excitation. The utility of optical recording has risen in recent years, but its signal-to-noise ratio is generally inferior to electrical measurements. Therefore, we believe the accumulation of data by both techniques will provide information about neuronal processing. Optical recording has the potential to elucidate neuronal networks in the hippocampus related to learning and memory, including LTP.

REFERENCES

1. 2. 3. 4. 5. 6. 7. 8.

Albowitz B., Kuhnt U. and Ehrenreich L. (1990) Optical recording of epileptiform voltage changes in the neocortical slice. Expl Brain Res. 81, 241–256. Amaral D. G. and Witter M. P. (1989) The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience 31, 571–579. Barish M. E., Ichikawa M., Tominaga T., Matsumoto G. and Iijima T. (1996) Enhanced fast synaptic transmission and a delayed depolarization induced by transient potassium current blockade in rat hippocampal slice as studied by optical recording. J. Neurosci. 16, 5672–5687. Cinelli A. R. and Salzberg B. M. (1992) Dendritic origin of late events in optical recordings from salamander olfactory bulb. J. Neurophysiol. 68, 786–806. Cohen L. B. and Lesher S. (1986) Optical monitoring of membrane of multisite optical measurement. In Optical Methods in Cell Physiology (eds De Weer P. J. and Salzberg B. M.), pp. 71–99. John Wiley and Sons, New York. Colom L. V. and Saggau P. (1994) Spontaneous interictal-like activity originates in multiple areas of the CA2–CA3 region of hippocampal slices. J. Neurophysiol. 71, 1574–1585. Grinvald A., Frostig R. D., Lieke E. and Hildesheim R. (1988) Optical imaging of neuronal activity. Physiol. Rev. 68, 1285–1366. Grinvald A., Manker A. and Segal M. (1982) Visualization of the spread of electrical activity in rat hippocampal slices by voltage-sensitive optical probes. J. Physiol. 333, 269–291.

8 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

Y. Nakagami et al. Halasy K. and Somogyi P. (1993) Subdivisions in the multiple GABAergic innervation of granule cells in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci. 5, 411–429. Han Z.-S., Buhl E. H., Lo¨rinczi Z. and Somogyi P. (1993) A high degree of spatial selectivity in the axonal and dendritic domains of physiologically identified local-circuit neurons in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci. 5, 395–410. Iijima T., Witter M. P., Ichikawa M., Tominaga T., Kajiwara R. and Matsumoto G. (1996) Entorhinal–hippocampal interactions revealed by real-time imaging. Science 272, 1176–1179. Ishizuka N., Weber J. and Amaral D. G. (1990) Organization of intrahippocampal projections originating from CA3 pyramidal cells in the rat. J. comp. Neurol. 295, 580–623. Jensen M. S. and Yaari Y. (1988) The relationship between interictal and ictal paroxysms in an in vitro model of focal hippocampal epilepsy. Ann. Neurol. 24, 591–598. Kamino K. (1991) Optical approaches to ontogeny of electrical activity and related functional organization during early heart development. Physiol. Rev. 71, 53–91. Konnerth A., Obaid A. L. and Salzberg B. M. (1987) Optical recording of electrical activity from parallel fibres and other cell types in skate cerebellar slices in vitro. J. Physiol. 393, 681–702. Laurberg S. and Sørensen K. E. (1981) Associational and commissural collaterals of neurons in the hippocampal formation (hilus fasciae dentatae and subfield CA3). Brain Res. 212, 287–300. Matsumoto G. and Ichikawa M. (1990) Optical system for real-time imaging of electrical activity with a 128#128 photopixel array. Soc. Neurosci. Abstr. 16, 490. Nakagami Y., Saito H. and Matsuki N. (1996) Optical recording of rat entorhino-hippocampal system in organotypic culture. Neurosci. Lett. 216, 211–213. Nicoll R. A. and Malenka R. C. (1995) Contrasting properties of two forms of long-term potentiation in the hippocampus. Nature 377, 115–118. Obaid A. L. and Salzberg B. M. (1996) Micromolar 4-aminopyridine enhances invasion of a vertebrate neurosecretory terminal arborization. J. gen. Physiol. 107, 353–368. Saggau P., Galvan M. and Bruggencate G. T. (1986) Long-term potentiation in guinea pig hippocampal slices monitored by optical recording of neuronal activity. Neurosci. Lett. 69, 53–58. Salzberg B. M. (1989) Optical recording of voltage changes in nerve terminals and in fine neuronal processes. A. Rev. Physiol. 51, 507–526. Sugitani M., Sugai T., Tanifuji M., Murase K. and Onoda N. (1994) Optical imaging of the in vitro guinea pig piriform cortex activity using a voltage-sensitive dye. Neurosci. Lett. 165, 215–218. Sugitani M., Sugai T., Tanifuji M. and Onoda N. (1994) Signal propagation from piriform cortex to the endopiriform nucleus in vitro revealed by optical imaging. Neurosci. Lett. 171, 175–178. Tamamaki N., Abe K. and Nojyo Y. (1987) Columnar organization in the subiculum formed by axon branches originating from single CA1 pyramidal neurons in the rat hippocampus. Brain Res. 412, 156–160. Tanifuji M., Sugiyama T. and Murase K. (1994) Horizontal propagation of excitation in rat visual cortical slices revealed by optical imaging. Science 266, 1057–1059. Vranesic I., Iijima T., Ichikawa M., Matsumoto G. and Kno¨pfel T. (1994) Signal transmission in the parallel fiber–Purkinje cell system visualized by high-resolution imaging. Proc. natn. Acad. Sci. U.S.A. 91, 13014–13017. Yeckel M. F. and Berger T. W. (1990) Feedforward excitation of the hippocampus by afferents from the entorhinal cortex: redefinition of the role of the trisynaptic pathway. Proc. natn. Acad. Sci. U.S.A. 87, 5832–5836. (Accepted 17 March 1997)