Changes in synaptic efficacy of dentate granule cells during operant behavior in rats

Physiology & Behavior 105 (2012) 938–947

Contents lists available at SciVerse ScienceDirect

Physiology & Behavior journal homepage: www.elsevier.com/locate/phb

Changes in synaptic efficacy of dentate granule cells during operant behavior in rats☆ Shigeki Nomoto a,⁎, Tomohiro Yamamoto b, 1, Jun-ichi Tomioka c, 2, Emi Nomoto a a b c

Department of Central Nervous System, Tokyo Metropolitan Institute of Gerontology, Itabashi-ku, Tokyo 173-0015, Japan Department of Biomolecular Science, Faculty of Science, Toho University, Funabashi, Chiba 274-8510, Japan Department of Regulation Biology, Faculty of Science, Saitama University, Saitama 338-8570, Japan

a r t i c l e

i n f o

Article history: Received 17 March 2010 Received in revised form 1 November 2011 Accepted 7 November 2011 Keywords: Operant behavior Synaptic efficacy Hippocampus Dentate gyrus Learning Memory Freely moving rat

a b s t r a c t It is widely accepted that long-term potentiation (LTP) is one of the fundamental physiological mechanisms underlying memory function based on its response properties and behavior of the induced sites. Many experimental approaches have been used to investigate whether the mechanisms underlying LTP are activated during learning and whether these mechanisms underlie the formation of certain types of memory. However, relatively few studies have reported the time course of changes in the efficacy of synaptic transmission in the learning process. We simultaneously monitored changes in slope of field EPSPs (fEPSP slope) and the amplitude of population spikes (pop. spike) in perforant path-evoked potentials in the dentate gyrus over the course of an appetitively motivated operant paradigm in freely moving rats. We found that the fEPSP slope recorded from the granule cell layer was potentiated about 7%, the fEPSP slope recorded from the molecular layer was depressed about 20%, and the amplitude of pop. spike recorded from the granule cell layer was significantly depressed about 40% after the trial in which rats began to press the lever frequently. These results suggested that the granule cells in the dentate gyrus received excitatory inputs in the somatic region and inhibitory inputs in the dendritic region, and that outputs from the granule cells were significantly reduced in the process of acquisition of the operant behavioral task. We observed no LTP in this study although our rats were capable of having LTP induced by a high-frequency stimulus. The depression of fEPSP slope induced without any artificial stimulation in this study is thought to be another neural mechanism underlying learning and memory. The origins of excitatory and inhibitory inputs are unknown at the moment. © 2011 Elsevier Inc. All rights reserved.

1. Introduction Long-term potentiation (LTP) was first observed in the dentate gyrus of the rabbit hippocampus [1,2]. Bliss and Collingridge stated that LTP of synaptic transmission in the hippocampus is the primary experimental model for investigation of the synaptic basis of learning and memory in vertebrates [3]. Many experimental approaches have been used to investigate whether the mechanisms underlying LTP are activated during learning and whether these mechanisms underlie Abbreviations: slopeTD, transient depression of fEPSP slope; slopeTP, transient potentiation of fEPSP slope; spikeTD, transient depression of pop spike. ☆ The present experimental protocol was reviewed and approved by the appropriate committee of the Tokyo Metropolitan Institute of Gerontology. We are also obligated to follow the Guiding Principles for Care and Use of Animals in the Field of Physiological Science of the Physiological Society of Japan. ⁎ Corresponding author at: Department of Central Nervous System, Tokyo Metropolitan Institute of Gerontology, 35-2 Sakaecho, Itabashi-ku, Tokyo 173-0015, Japan. Tel.: + 81 3 3964 3241x3084; fax: + 81 3 3579 4776. E-mail address: [email protected] (S. Nomoto). 1 Present address: Kissei Pharmaceutical Co., Ltd., Matsumoto, Nagano 399-8710, Japan. 2 Present address: Japan Automobile Research Institute, Higashi-Ibaraki-gun, Ibaraki 311-4316, Japan. 0031-9384/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.physbeh.2011.11.010

the formation of certain types of memory. However, relatively few studies have reported the time course of changes in the efficacy of synaptic transmission in the learning process. One means of addressing this question is to monitor synaptic field potentials over the time course of learning [4]. Weisz et al. (1984) examined the responsivity of dentate gyrus granule cells to perforant path stimulation during classical conditioning of the rabbit nictitating membrane response [5]. They demonstrated above-baseline increases in dentate population spike amplitudes over the course of training in paired but not unpaired animals. However, they also observed that population spike amplitudes were smaller when elicited during tone presentations in both paired and unpaired animals than between trials when no conditioning stimuli were present. Doyère et al. measured changes in perforant path-dentate gyrus-evoked field potentials in rats performing a classical conditioning (paired tone and footshock) or pseudoconditioning (unpaired tone and footshock) task to determine the time course of synaptic modification during learning [4]. They demonstrated an increase in the slope of the field excitatory postsynaptic potential (fEPSP) in the conditioned group and a decrease in the slope of the fEPSP in the pseudoconditioning group. They concluded that behavioral events can exert bi-directional control of synaptic strength of entorhinal cortex inputs to the dentate gyrus and that

S. Nomoto et al. / Physiology & Behavior 105 (2012) 938–947

939

learning-induced changes in the efficacy of synaptic transmission. We simultaneously monitored changes in fEPSP slopes and the amplitude of population spikes in perforant path-evoked potentials in the dentate gyrus over the course of an appetitively motivated operant paradigm in freely moving rats.

the sign of synaptic modification is at least in part determined by the temporal relationship between these events. Dentate population spike amplitudes and fEPSP slopes are thus quite variable under different behavioral conditions. Matthies et al. used stimulation of the perforant path with impulse trains of 15 cps and 670 ms duration as a conditioned stimulus in a two-way shuttle box avoidance in rats. Field potentials in the dentate area evoked by test stimuli were measured after training sessions until the 7th day. They reported that foot-shock and unconditioned escape elicited only a transient slight depression of the amplitude of population spikes and also slightly increased the slope of the population EPSP of evoked test potentials. Control stimulation of the perforant path without pairing with foot-shock, as in conditioning, only slightly increased the slope of the population EPSP of evoked test potentials, but produced strong transient inhibition followed by long-lasting moderate depression of the population spike amplitude [6]. In experiments by Skelton et al., the efficacy of synaptic transmission from the perforant path (PP) to the granule cells in the dentate gyrus (DG) of freely moving rats was monitored electrophysiologically over the course of training in an appetitively motivated, discriminated operant paradigm. They recorded evoked potentials from the DG following stimulation of the PP and measured the amplitudes of population spikes. They found that significant increases in population spike amplitudes over 8 days of training but not over 8 days of free feeding, and suggested that such increases in synaptic efficacy may encode some aspect of learning [7]. To our knowledge, there have been no reported attempts to measure the efficacy of synaptic transmission from the perforant path to the cell and molecular layers of granule cells in the dentate gyrus in the process of acquisition of an appetitively motivated operant conditioning paradigm. We therefore examined potential

2. Materials and methods The present experimental protocol was reviewed and approved by the appropriate committee of the Tokyo Metropolitan Institute of Gerontology. We are also obligated to follow the Guiding Principles for Care and Use of Animals in the Field of Physiological Science of the Physiological Society of Japan. 2.1. Animals Male Fischer 344 rats weighing 318–419 g were used as subjects. Two or three rats were housed in the same cage with water and food available ad libitum. Rats were kept in a temperaturecontrolled room maintained at 23 ± 1 °C with a 12-h light–dark cycle (lights on at 08.00 h). Domesticated rats were handled for 5–10 min/day several times per week until the end of the experiment. 2.2. Surgery and electrophysiology Animals were prepared for chronic recording and stimulation 1 week after the start of handling. The surgery was carried out under pentobarbital sodium anesthesia (40 mg/kg, i.p.), and core temperature was maintained at 37 ± 0.5 °C during the operation. Monopolar teflon-coated stainless steel recording and stimulating

Hipp fiss

A

B

Rec.

CA1

Stim.

S CA3

Molecular layer

pp

Granule cell layer

gc

sc mf

pp

DG

C

mf

D

E

2 mV

2 mV 10 ms

10 ms

Fig. 1. A, schema of the electrophysiological arrangement. The perforant pathway was stimulated with a biphasic square-wave constant current pulse of 0.5 ms (full width) every 30 s. Field potentials were recorded from the dentate area (molecular layer or granule cell layer). B, the region enclosed in the rectangle in A enlarged to show the apical dendritic field of the granule cells, with the perforant path fibers confined to the central one-third of the field. C, the field potential evoked by a perforant path stimulus recorded in the molecular layer. Slope of field EPSP was measured as maximal slope (solid line) of falling phase. D, the field potential evoked by a perforant path stimulus recorded in the granule cell layer. Slope of field EPSP was measured as maximal slope (solid line) of rising phase. E, amplitude of population spikes was measured as vertical distance (solid line) from the bottom to tangent line (broken line) between two peaks. Abbreviations: DG, dentate gyrus; CA1, CA3, fields of hippocampus; S, subiculum; pp, perforant path fibers from the entorhinal cortex; mf, mossy fibers from the granule cells; sc, Schaffer collateral connections from CA3 to CA1; gc, granule cells; Hipp fiss, hippocampal fissure.

940

S. Nomoto et al. / Physiology & Behavior 105 (2012) 938–947

electrodes (75 μm bare, 113 μm coated; A-M Systems, Everett, WA, USA) were positioned in the dentate gyrus and perforant pathway of the left hemisphere, respectively (Fig. 1A). The recording electrode was inserted 3.4 mm posterior to the bregma and 2.0 mm left of the midline, and the stimulating electrode was inserted 8.0 mm posterior to the bregma and 3.8 mm left of the midline. The recording electrode was inserted into the molecular layer (approximately 2.5 mm depth) or into the granule cell layer (approximately 2.8 mm depth) in different animals (Fig. 1B). Optimal placement of the electrodes in the dentate gyrus was determined using the laminar profile of field potentials (Fig. 1C–E) elicited by pulses to the ipsilateral perforant pathway as a guide (Fig. 2), and was verified by postmortem examination (Figs. 3, 4). The electrodes were fixed permanently to the skull with dental cement after increasingly negative or positive field potentials to stimulation had been maximized. At least 7 days were allowed for recovery from surgery. The field potentials were evoked by stimulation with a biphasic square-wave constant current pulse (150–500 μA) of 0.5 ms duration (full width) every 30 s throughout the operant behavior testing period of 60 min. At the beginning of each experiment, input/output curves were determined to ascertain the maximum evoked fEPSP slope. For measurement of basal synaptic transmission, stimulus intensity was used, which evoked a response that was 50–60% of the maximum. 2.3. Measurement of the slope of fEPSPs and the amplitude of population spikes As shown in Fig. 1C, the field potentials evoked by perforant path stimuli were recorded in the molecular layer. The slope of fEPSPs was measured as maximal slope (solid line) in the falling phase. As shown in Fig. 1D and E, the field potentials evoked by perforant path stimuli were recorded in the granule cell layer. The slope of fEPSPs was measured as maximal slope (solid line) in the rising phase (Fig. 1D). The amplitude of population spikes was measured as vertical distance (solid line) from the bottom to the tangent line (broken line) between two peaks (Fig. 1E). 2.4. Operant behavioral task After the recovery period following surgery, food was progressively reduced until each rat reached 75–85% of its free-feeding body weight during the 7-day period. Rats were weighed daily throughout the

2

3

4

0

1

2

3

4 mm

Bregma-3.30 mm Fig. 3. Schematic drawing of recording electrode placements in the molecular layer of the dentate gyrus. Shown is a coronal view at position of 3.30 mm posterior to the bregma. Solid gray circles indicate the locations.

experimental period, and the diet was adjusted for each rat in order to maintain homogeneous weight loss. For behavioral testing, we chose a positive reinforcement behavior test using an operant chamber (ENV-007CT, MED Associates, Georgia, VT, USA). The chamber was enclosed in a sound-attenuating cubicle (56 cm width, 35.5 cm length, 56 cm height) with a ventilating fan. The inner dimensions of the operant chamber were 24 cm in width, 30.5 cm in length, and 29 cm in height, with a grid floor of stainless steel bars. A response lever (LV) was mounted 3 cm above the floor and 4 cm from the left wall. The lever required a force of 10 g for operation. A food dispenser (PR) for food pellets (Dustless precision pellets 45 mg, F0021, Bio-Serve, Frenchtown, NJ, USA) was located 4 cm to the right of the lever and 2 cm above the floor. A key stimulus lamp (KL) was located 8 cm above the lever (Fig. 5). A house light was located in the back wall 28 cm from the floor. Apparatus control and data sampling were achieved with behavioral programming (MED-PC Ver. 2.08, MED Associates, Georgia, VT, USA). We connected lead wires to the stimulating and recording electrodes and placed the rat in the operant chamber and then started recording field potentials for 1 h. The rat could move freely in the chamber. During the operant behavior task, the perforant path was stimulated every 30 s and field potentials were recorded from the molecular layer or granule cell layer of the dentate gyrus. The rat's behavior was monitored continuously with a video system from outside of the sound-attenuating cubicle.

1 2 2

3 3 4

5

0

1

2

3

4

5

6

7

4

mm

Bregma-8.00 mm

0

1

2

3

4 mm

Bregma-3.30 mm Fig. 2. Schematic drawings of stimulating electrode placements. After the completion of the experiment, the rats were given marking lesions of the perforant path. Shown is a coronal view at position 8.00 mm posterior to the bregma. Solid gray circles indicate the locations.

Fig. 4. Schematic drawing of recording electrode placements in the granule cell layer of the dentate gyrus. Shown is a coronal view at position of 3.30 mm posterior to the bregma. Solid gray circles indicate the locations.

S. Nomoto et al. / Physiology & Behavior 105 (2012) 938–947

Key Lamp:

OFF

ON

OFF-1 0

Food Pellets:

OFF

ON

OFF-2

15

not provided

Electric stimulus: 0 (every 30 s)

1

2

941

45

3

Key Lamp OFF

60 min

not provided

provided

57 58 59 60 min

Key Lamp ON

KL KL

KL KL

PR PR

LV LV

PR PR

LV LV

Fig. 5. Procedure of the operant behavior task over 1 h. The key lamp (KL) was off for the first 15 min (OFF-1). During this period, rats were not provided a food pellet even if they pressed the response lever (LV). During the following 30 min (ON), the KL was on and rats were provided food pellets in the pellet receptacle (PR) when they pressed the LV. During the following 15 min (OFF-2), the KL was off and rats were not provided with a food pellet even if they pressed the lever. During the operant behavior task, the perforant path was stimulated every 30 s and field potentials were recorded from the molecular layer or granule cell layer of the dentate gyrus.

The animals were placed into the chamber only once per day. The operant behavioral test lasted 1 h (Fig. 5). During the first 15 min, the key stimulus lamp was off (Phase OFF-1); during the next 30 min, the lamp was on (Phase ON); and during the last 15 min, it was off (Phase OFF-2). During Phase ON, one press of the lever was quickly rewarded with one food pellet. When each rat pressed the lever more than 150 times during Phase ON, the schedule was altered so that three lever pressings were quickly rewarded with one food pellet in subsequent experiments to prevent reduction of appetitive motivation. During Phases OFF-1 and OFF-2, no food pellet was supplied even if the rat pressed the lever. In the process of acquisition of the operant behavioral task, the frequency of lever-pressing dramatically changed. No rats pressed the lever in purposeful fashion at the beginning of the behavioral task, although rats did touch the lever accidentally with their bodies. However, once rats noticed the relationship between lever-pressing and appearance of a food pellet, they began to press the lever in purposeful fashion with their forelimb or muzzle (as observed by video). In this paradigm, responses that initially resulted from random searches of the food dispenser were shaped into precise, short-latency responses for food pellets. Each rat required a different number of trials to acquire the operant behavioral task. One rat required only two trials to acquire the paradigm, while two rats required 23 trials. We therefore defined “the first trial of acquisition” based on the frequency of purposeful lever-pressing by the forelimb or muzzle increasing more than 15 times during 30 min of Phase ON. We rearranged the data for each rat at the first trial of acquisition and carried out statistical analysis. With this method of processing of data, trials before the first trial of acquisition were assumed to be training sessions, and the total period of training sessions thus varied among rats in this experiment. We also defined achievement of memory formation of the operant behavioral task as the frequency of lever-pressing during 30 min of Phase ON reaching 80% level of the maximum frequency of each rat. Such rats continued to press the lever even though food pellets were not provided in the following 2 or 3 trials (unpublished data).

The rats were continuously observed with a video camera (CCD-MC1, SONY, Tokyo, Japan), and a video recorder (CCD-TRV92, SONY, Tokyo, Japan) was used to store the images on videotapes throughout the operant test. 2.5. Data analysis The mean slope of fEPSP in Phase OFF-1 was set at 100%, and data during Phase ON and Phase OFF-2 were expressed as percent changes (Fig. 2B). All data are presented as means ± SEM. Statistical analysis was carried out using standard two-way parametric ANOVA coupled with Bonferroni/Dunn fractal ANOVA for repeated comparisons. Values of P b 0.05 were considered significant. 3. Results 3.1. Operant behavioral task and field potentials recorded from the molecular layer of the granule cells Fig. 6 shows a typical example of the frequency of lever-pressing (A), the original waveforms of field potentials recorded from the molecular layer of the granule cells (B), and the slope of fEPSPs every 30 s (C). These were recorded in the seventh trial after acquisition in one rat. The small number of lever presses in phase OFF-1 was a result of the expectation of obtaining food pellets, though no food pellets were rewarded in this phase (Fig. 6A). After the key stimulus lamp was turned on (Phase ON), the number of lever presses rapidly increased and the rat was rewarded with a food pellet every three lever presses. During this phase, the number of lever presses gradually decreased from 210 to 130 times/5 min. The frequency of lever-pressing decreased after the lamp was turned off (phase OFF-2), and no food pellet was rewarded during this phase. fEPSP slopes (mV/ms) were measured every 30 s (open circles in Fig. 6C). Mean fEPSP slopes were calculated in each phase of OFF-1, ON, and OFF-2 (black filled circles in Fig. 6C, −0.612 ± 0.015, − 0.387 ± 0.009, −0.394 ± 0.011 mV/ms, respectively). In Fig. 6B, original

942

S. Nomoto et al. / Physiology & Behavior 105 (2012) 938–947

Frequency of lever-pressing (times)

A 250

OFF-1

ON

OFF-2

200 150 100 50 0

Time (min)

B

C

Slope of fEPSP (mV/ms)

(1)

(2)

(3)

0

OFF-1

-0.1

ON

OFF-2 (3)

-0.2

(2)

-0.3 -0.4 -0.5

(1)

-0.6 -0.7 -0.8 0

5

10

15

20

25

30

35

40

45

50

55

60

Time (min) Fig. 6. Example of frequency of lever pressing (A), the original waveforms of field potentials recorded from the molecular layer at each point (1, 2 and 3) indicated as gray filled circles in C (B), and slope of fEPSP (C), recorded in the seventh trial after acquisition in freely moving Rat-136. Black filled circles in C are mean values in each phase. Calibrations in B: 5 ms and 1 mV.

waveforms of the field potentials are shown at each point (gray filled circles) indicated in Fig. 6C. The slope of fEPSP decreased slightly during phase OFF-1 after the rat had been caged in the operant chamber. After the key stimulus lamp was turned on, the fEPSP slope decreased further during phase ON. Then the fEPSP slope increased gradually during phase OFF-2 after the lamp was turned off (Fig. 6C). Fig. 7 shows other examples from the fourth trial before acquisition (n = 7, Fig. 7A,B) and the second trial after acquisition, in which rats purposefully pressed the lever with the forelimb or muzzle and the frequency of lever-pressing increased more than 15 times during Phase ON (n = 8, Fig. 7C,D). Rats did not press the lever frequently before acquisition. The small numbers of lever presses in the first 20–25, 25–30, and 55–60 min were the result of rats unintentionally bumping against the lever with their bodies (Fig. 7A). Slopes of fEPSP (please note: expressed in %) recorded from the molecular layer were measured every 30 s (open circles in Fig. 7B,D). Mean fEPSP slopes were calculated in each phase of OFF-1, ON, and OFF-2 (filled circles in Fig. 7B,D). The fEPSP slope of field potentials did not significantly change during the 1-h experiment before acquisition (Fig. 7B, Mean fEPSP slope ± SEM: 100.02 ± 1.21%, 97.77 ± 0.74%, and 96.63 ± 1.07%, respectively). However, once rats acquired the operant paradigm, they began to press the lever frequently. The rats pressed the lever about 50 times in total during Phase OFF-1, but no food pellet was rewarded in this phase. When the key stimulus lamp was on (Phase ON), the frequency of lever-pressing rapidly increased and rats were rewarded

a food pellet with every three lever-pressings, but then gradually decreased in this phase. The frequency of lever-pressing decreased after the lamp was turned off (Phase OFF-2) and no food pellet was rewarded (Fig. 7C). After the rats were caged in the operant chamber, the fEPSP slope either remained unchanged or sometimes slightly decreased during Phase OFF-1 (Fig. 7D). When the key stimulus lamp was on, the rats began to press the lever more frequently, and the fEPSP slope slowly began to decrease during Phase ON (Mean fEPSP slope ± SEM: 91.86 ± 0.75%, F(2,117) = 28.30, P b 0.0001 vs. OFF-1; Bonferroni/Dunn test after one-way ANOVA). After the lamp was turned off, the fEPSP slope either gradually increased or sometimes remained unchanged during Phase OFF-2 (Mean fEPSP slope ± SEM: 89.59 ± 1.04%, F(2,117) = 28.30, P b 0.0001 vs. OFF-1; Bonferroni/Dunn test after one-way ANOVA). Each rat required a different number of trials to acquire the operant behavioral task. We therefore defined “the first trial of acquisition” by the frequency of positive lever-pressing by the forelimb or muzzle increasing more than 15 times during 30 min of Phase ON. We rearranged the data for each rat at the first trial of acquisition and performed statistical analysis (for further details, see Materials and Methods 2.4). Fig. 8A shows the frequency of lever-pressing during the process of acquisition of operant behavior (n = 7). The rats did not press the lever until the first trial. However, once they pressed the lever and received food pellets, they began to press it frequently. The frequency of lever-pressing increased at an increasing tempo and reached a level of approximately 600 times during Phase ON in the 7th trial (filled circles). Rats pressed the lever not only during Phase ON, but also during Phase OFF-1 (filled triangles) and Phase OFF-2 (reverse open triangles). Rats were rewarded food pellets only during Phase ON, but not during Phase OFF-1 or Phase OFF-2. The frequency of lever-pressing during Phase OFF-2 was lower than that during Phase OFF-1. The frequency of lever-pressing during Phase OFF-2 decreased gradually as the experiment proceeded, although the frequency of lever-pressing during Phase OFF-1 remained nearly constant. Open circles indicate total frequency of lever-pressing during a 1-h experiment. fEPSPs were recorded from the molecular layer. In each trial, the mean fEPSP slope during Phase OFF-1 was set at 100% (filled triangles), and the mean fEPSP slopes during Phase ON and Phase OFF-2 were expressed as percentages. As shown in Fig. 8B, the fEPSP slope during Phase ON (filled circles) was significantly decreased to approximately 80% compared with the mean values for Phase OFF-1. These significant decreases were observed in the 3 rd, 8th, 17th, 19th, 28th, and 30th trials after the rats began to press the lever frequently (n = 8, *: P b 0.05). The fEPSP slope during Phase OFF-2 (reverse open triangles) changed in the same fashion as that during Phase ON. A significant decrease was observed in the 30th trial (n = 8, †: P b 0.05). The depression lasted for 45 min or more after the key stimulus lamp was turned on in each experiment. We did not record field potentials after the trial, but the depression disappeared until the next trial approximately 24 h after the last trial. We therefore termed the depression “transient depression” (slopeTD). These changes in the fEPSP slope recorded from the molecular layer could not be observed before rats acquired the operant paradigm of pressing the lever. SlopeTD in the perforant path-evoked dentate granule-cell response recorded from the molecular layer was strongly correlated with the acquisition of operant behavior. SlopeTD appeared to be induced periodically after the acquisition of operant behavior until the 30th trial. 3.2. Operant behavioral task and field potentials recorded from the granule cell layer This experiment was performed in animals different from those used in Experiment 3.1.

A 5

ON

OFF-1

OFF-2

4 3 2 1 0 0-5 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50 50-55 55-60

Time (min) 120

Frequency of lever-pressing (times)

Frequency of lever-pressing (times)

S. Nomoto et al. / Physiology & Behavior 105 (2012) 938–947

B

ON

OFF-1

OFF-2

0-5 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50 50-55 55-60

Time (min)

D

110

Slope of fEPSP (%)

Slope of fEPSP (%)

C 50 45 40 35 30 25 20 15 10 5 0

120

110

943

100 90 80 70 60

100 90 80 70 60

ON

OFF-1

50 0

OFF-1

OFF-2

OFF-2

ON

50 15

30

45

60

Time (min)

0

15

30

45

60

Time (min)

Fig. 7. Changes in the frequency of lever-pressing and in the fEPSP slope from the experiments before (-4th trial, n = 7, A and B) and after (2nd trial, n = 8, C and D) the acquisition of operant behavior. fEPSPs were recorded from the molecular layer. Black filled circles in B and D are mean values in each phase.

Frequency of lever-pressing (times)

A

1000 900 800 700 600 500 400 300 200 100 0 -15

-10

-5

0

5

10

15

20

25

30

Trial

Slope of fEPSP (%)

120

B

110 100

* 90 80

* *

70 60 -15

-10

-5

0

* *

5

10

15

20

* 25

30

Trial Fig. 8. Changes in the frequency of lever-pressing and in the fEPSP slope during operant behavior in freely moving rats. A, Frequency of lever-pressing (open circles: total frequency of lever-pressing during a 1-h experiment, filled circles: frequency during Phase ON, filled triangles: frequency during Phase OFF-1, reverse open triangles: frequency during Phase OFF-2). B, In each trial, the mean slope of the fEPSP, recorded from the molecular layer, during Phase OFF-1 was set at 100% (filled triangles), and the mean fEPSP slopes during Phase ON and Phase OFF-2 are expressed as percentages (filled circles and reverse open triangles, respectively). Means ± SEM.

Fig. 9 shows a typical example of the frequency of lever-pressing (A), the original waveforms of field potentials recorded from the granule cell layer at each point (1, 2 and 3) indicated as gray filled circles in C and D (B), and slope of fEPSP (C), and amplitude of population spike (D), recorded in the sixth trial after acquisition in freely moving Rat-352. Fig. 10A shows the frequency of lever-pressing during the process of acquisition of operant behavior (n = 8). The frequency of leverpressing increased day by day after the acquisition of operant behavior (trials 1 to 6), and exceeded 600 times during Phase ON in the 5th trial period (filled circles). fEPSPs and population spikes could be recorded from the granule cell layer. In each trial, the mean fEPSP slope during Phase OFF-1 was set at 100% (filled triangles), and the mean fEPSP slopes during Phase ON and Phase OFF-2 were expressed as percentages. As shown in Fig. 10B, the fEPSP slope tended to decrease slightly during phase OFF-2 before the rat acquired the operant behavior task (trials -3 to -1). However, the fEPSP slope during Phase ON (filled circles) was increased to 107% after the trial when the rats started to press the lever frequently (F(1,52) = 22.53, P b 0.0001 vs. OFF-1) (trials 1 to 6). The fEPSP slope during Phase OFF-2 (reverse open triangles) changed in the same fashion as that during Phase ON (F(1,52) = 49.40, P b 0.0001 vs. OFF-1). During Phase ON, significant increases in fEPSP slope, compared with the mean values of Phase OFF-1, were observed in the 3 rd and 6th trials (n = 6, *: P b 0.05). During Phase OFF-2, a significant decrease and increase in fEPSP slope, compared with the mean values of Phase OFF-1, were observed in the -1st and 3 rd trials, respectively (n = 6, †: P b 0.05). The potentiation lasted for 45 min or longer after the key stimulus lamp was turned on in each experiment. The potentiation disappeared until the next trial approximately 24 h after the last trial. Therefore, we termed the potentiation transient potentiation (slopeTP).

944

S. Nomoto et al. / Physiology & Behavior 105 (2012) 938–947

180 160 140 120 100 80 60 40 20 0

OFF-1

0

5

10

ON

15

20

25

30

Frequency of lever-pressing (times)

Frequency of lever-pressing (times)

A OFF-2

35

40

45

50

55

60

1000 900 800 700 600 500 400 300 200 100 0

A

-3

Time (min)

(3)

8

ON (2)

6

(3)

2 0 5

10

15

20

25

30

35

40

45

50

55

60

Amplitude of pop. Spike (mV)

14

OFF-1

12

1

2

3

4

5

6

B *

110 *

100 90 80 70 60 -2

-1

0

1

ON

2

3

4

5

6

Trial

4

0

0

OFF-2

Time (min)

D

120

-3

OFF-1 (1)

10

Amplitude of pop. Spike (%)

Slope of fEPSP (mV/msec)

10 mV

Slope of fEPSP (%)

(2)

10 msec

C

-1

Trial

B (1)

-2

100 90 80 70 60 50 40 30 20 10 0

C * *

*

*

*

*

*

*

-3

OFF-2

*

*

-2

-1

0

1

2

3

4

5

6

Trial

10 8

(1)

6 4

(2)

2

(3)

0 0

5

10

15

20

25

30

35

40

45

50

55

60

Time (min) Fig. 9. Example of frequency of lever pressing (A), the original waveforms of field potentials recorded from the granule cell layer at each point (1, 2 and 3) indicated as gray filled circles in C and D (B), slope of fEPSP (C), and amplitude of population spike (D), recorded in the sixth trial after acquisition in freely moving Rat-352. Black filled circles in C and D are mean values in each phase.

On the other hand, the amplitude of population spikes (pop. spike) decreased significantly during Phase ON and OFF-2 compared to OFF-1 throughout the experiment (ON: F(1,61) = 34.17, P b 0.0001 vs. OFF-1 and OFF-2: F(1,61) = 38.50, P b 0.0001 vs. OFF-1). During Phase ON and OFF-2, significant decreases, compared with the mean values of Phase OFF-1, were observed throughout the experiment (n = 7, *: P b 0.05 for Phase ON and †: P b 0.05 for Phase-2). The depression in pop. spike amplitude reached approximately 60% after the rats started to press the lever frequently (trials 2 to 6) (Fig. 10C). The depression lasted for 45 min or longer after the key stimulus lamp was turned on in each experiment. We did not record field potentials after the trial, but the depression disappeared until the next trial approximately 24 h after the last trial. We therefore termed the depression transient depression of pop. spike (spikeTD).

Fig. 10. Changes in the frequency of lever-pressing and in the fEPSP slope during operant behavior in freely moving rats. A, Frequency of lever-pressing (open circles: total frequency of lever-pressing during a 1-h experiment, filled circles: frequency during Phase ON, filled triangles: frequency during Phase OFF-1, reverse open triangles: frequency during Phase OFF-2). B, In each trial, the mean slope of the fEPSP, recorded from the granule cell layer during Phase OFF-1 was set at 100% (filled triangles), and the mean fEPSP slopes during Phase ON and Phase OFF-2 are expressed as percentages (filled circles and reverse open triangles, respectively). C, In each trial, the mean amplitude of pop. Spike, recorded from the granule cell layer during Phase OFF-1 was set at 100% (filled triangles), and the mean amplitudes during Phase ON and Phase OFF-2 are expressed as percentages (filled circles and reverse open triangles, respectively). Means ± SEM.

These changes in the fEPSP slope and pop. spike amplitude could not be observed before rats acquired the operant paradigm of pressing the lever. SlopeTP and spikeTD in the perforant path-evoked dentate granule-cell responses recorded from the granule cell layer were strongly correlated with the acquisition of operant behavior. 4. Discussion Electrical field recordings represent the summed responses from a number of neurons in the vicinity of the recording electrode. The waveform recorded from the molecular layer (Fig. 1C) is termed a field excitatory postsynaptic potential, or fEPSP, to indicate that the measured potential results from the summed activity across a population of granule cells. The current flowing into the dendrites during this fEPSP will exit the neurons near the cell body layer so that a field electrode in the granule cell layer will record a positivegoing potential during this same synaptic event (Fig. 1D). If the intensity of synaptic input is sufficient to evoke action potentials in the

S. Nomoto et al. / Physiology & Behavior 105 (2012) 938–947

neurons, then the field electrode in the granular layer will also record a negative-going potential (termed a population spike, Fig. 1E) resulting from the inward current during the postsynaptic action potentials. Measurements of the initial slope of the fEPSP measured in the molecular layer or granule cell layer provide a reliable estimate of the intensity of synaptic activity, whereas a measure of the amplitude of the population spike provides an estimate of the number of neurons reaching threshold from this synaptic input [8]. In this study, we found that the fEPSP slope recorded from the granule cell layer was potentiated about 7% (slopeTP), the fEPSP slope recorded from the molecular layer was depressed about 20% (slopeTD), and the amplitude of pop. spike recorded from the granule cell layer was significantly depressed about 40% (spikeTD) after the trial in which rats began to press the lever frequently. These results suggested that the granule cells in the dentate gyrus received excitatory inputs in the somatic region and inhibitory inputs in the dendritic region, and that outputs from the granule cells were significantly reduced in the process of acquisition of the operant behavioral task. Doyère et al. have indicated the difficulty in detecting changes in synaptic potentials during learning, and have mentioned that this difficulty can generally be attributed to two possible sources [4]. First, learning that induces synaptic changes over a limited set of active afferents might be difficult to detect with the field potential recording technique. Second, potentiation of one set of synapses may be accompanied by a depression at other synapses, making averaged synaptic strength constant. However, some authors have reported that the slopes of fEPSP increased in classical conditioning (paired tone and footshock) [4], and in the other conditioning with foot-shock, unconditional escape in a two-way shuttle box avoidance, and control stimulation of the perforant path without pairing with foot-shock in rats [6]. These results are similar to those of our fEPSP recording in the granule cell layer, but not those in the molecular layer. Other authors have reported that the amplitudes of dentate population spikes increased during conditioning of the nictitating membrane response in rabbits [5] and during an appetitively motivated, discriminated operant paradigm in rats [7]. These findings are opposite those of the present study. Matthies et al. also reported that foot-shock and unconditioned escape elicited only a transient slight depression in the amplitude of population spikes and also slightly increased the slope of the population EPSPs of the evoked test potentials. Control stimulation of the perforant path, without pairing with foot-shock as in conditioning, only slightly increased the slope function of test potentials, but produced a strong transient inhibition followed by a long-lasting moderate depression of the population spike amplitude [6]. These results are similar to ours recorded from the granule cell layer in the process of acquisition of the operant behavioral task, although our experimental methodology differed markedly from theirs. Skelton et al. found significant increases in population spike amplitudes over 8 days of operant conditioning, but not over 8 days of free-feeding [7]. In contrast, we recorded significant decreases in population spike amplitude in the process of acquisition of the operant behavioral task. Unfortunately, they did not record any slopes of fEPSPs of evoked potentials in their study. There is thus no electrophysiological consensus regarding the efficacy of synaptic transmission from the perforant path to the granule cells in the dentate gyrus during and/or after learning behavior. The only point of correspondence is thus the increase in fEPSP slope in the conditioning experiments [4,6] and in the present study where fEPSP slope was measured from the granule cell layer. Bliss and Lømo [1] found and other authors confirmed that alterations of the fEPSP slope and the amplitude of pop. spike in the course of long-term potentiation are not strongly correlated, increases in the amplitude of pop. spike occur without changes in the fEPSP slope, or the amplitude of pop. spikes may increase much

945

more strongly than the fEPSP slope [9]. We also observed an inverse correlation in this study in which decreases in the amplitude of pop. spike occurred with slight increases in the fEPSP slope in the granule cell layer. However, we observed decreases in the fEPSP slope in the molecular layer in the same process of acquisition of the operant behavioral task. Granule cells in the dentate gyrus receive many types of afferent input from intrinsic and extrinsic hippocampal neurons. Much of the sensory information that reaches the hippocampus enters through the entorhinal cortex. The major input to the dentate gyrus is from the entorhinal cortex, and the dentate gyrus receives no direct inputs from other cortical structures. Neurons located in layer II of the entorhinal cortex give rise to a pathway, the perforant path, that terminates both in the dentate gyrus and in the CA3 field of the hippocampus. Cells in the medial entorhinal cortex contribute axons that terminate in a highly restricted fashion within the middle portion of the molecular layer of the dentate gyrus, and those from the lateral entorhinal cortex terminate in the outer third of the molecular layer [8]. The inner third of the molecular layer of the dentate gyrus receives a projection that originates exclusively from cells in the polymorphic layer [10,11]. Because this projection originates on both the ipsilateral and contralateral sides, it has been called the ipsilateral associational-commissural projection. The ipsilateral associational and commissural projections appear to originate as collaterals from axons of the mossy cells of the hilus [11]. Most terminals of this pathway form asymmetrical, presumably excitatory synaptic terminals on spines of the granule cell dendrites [12,13]. Because the mossy cells are immunoreactive for glutamate [14], it is likely that they release this excitatory transmitter substance at their terminals within the ipsilateral associational-commissural zone of the molecular layer. In the present study, we observed slight increases in the fEPSP slope in the granule cell layer in the process of acquisition of the operant behavioral task. Therefore, granule cells appear to receive excitatory inputs from the entorhinal afferents and associational and commissural afferents in the molecular layer near the somata. There are a variety of basket cells located close to the granule cell layer. These all appear to contribute to the very dense terminal plexus that is confined to the granule cell layer. The terminals in this basket plexus are GABAergic and form symmetric, presumably inhibitory, contacts primarily on the cell body and shafts of apical dendrites of the granule cells [15]. A second, inhibitory input to granule cells originates from the axon–axonic or “chandelier-type” cells located in the molecular layer [16,17]. Another intrinsic projection within the dentate gyrus arises from a population of somatostatinimmunoreactive neurons scattered throughout the polymorphic layer [18,19]. These somatostatin cells, located in the polymorphic layer, colocalize with GABA and contribute a plexus of fibers and terminals to the outer portions of the molecular layer. This system of fibers, which forms contacts on the distal dendrites of the granule cells, provides a third means for inhibitory control over granule cell activity [20]. In the present study, we observed decreases in the fEPSP slope at the molecular cell layer in the process of acquisition of the operant behavioral task. Granule cells thus appear to receive inhibitory inputs from neurons such as basket cells, chandelier-type cells, and somatostatin-immunoreactive neurons to the outer potions of the molecular layer. The dentate gyrus receives subcortical inputs which originate mainly from the septal nuclei, from the supramamillary region of the posterior hypothalamus, and from several monoaminergic nuclei in the brainstem, especially the locus coeruleus and the raphe nuclei [21]. On the other hand, the amygdala is involved in memory via its mediation of emotional experiences, based on the idea that emotional attributes influence memory [22-26]. Akirav and Richter-Levin have proposed that the activation of the basolateral group of the amygdala (either by behavioral stress or by direct electrical

946

S. Nomoto et al. / Physiology & Behavior 105 (2012) 938–947

stimulation) has a biphasic effect on hippocampal plasticity that includes an immediate excitatory effect and a longer-lasting inhibitory effect [27]. The anterior cingulate cortex has anatomically strong connections with other limbic areas, such as the amygdala, mediodorsal thalamic nuclei, hypothalamus, hippocampal formation, and orbital cortex, all of which are implicated in emotion and behavioral control [28-34]. Many early studies suggested that the cingulate cortex was related to emotion and noxious stimuli [35,36]. The cingulate gyrus is also involved in behavior [37] and learning [38]. Lesion of the anterior cingulate cortex induced changes in emotion or anxiety-related behaviors in rats [39-41]. Takenouchi et al. (1999) recorded neuronal activity from the anterior cingulate cortex of behaving rats during discrimination and learning of conditioned stimuli with or without associated reinforcements. They also reported that stimulus attributes predicting reward or no reward are represented in the rostral and ventral parts of the anterior cingulate cortex, while the caudal and dorsal parts of the anterior cingulate cortex are related to execution of learned instrumental behaviors [42]. Thus, changes in the efficacy of the granule cells in the present study might also have been affected by these subcortical inputs. However, the involvement of these neurons must be examined in detail in further study. There appear to be three possible mechanisms underlying the slopeTD in the dentate gyrus at the beginning of the acquisition of operant behavior. A reduction of sensory (visual, auditory, tactile, etc.) information from the entorhinal cortex to the hippocampus through the perforant path is one possibility. Iijima et al. have reported in a slice experiment that the entorhinal neuronal circuit can contribute to memory processes by holding sensory information and selectively gating the entry of information into the hippocampus [43]. This mechanism may indeed be in part responsible for the slopeTD during the acquisition of operant behavior in freely moving rats. Second, the effect of inhibitory presynaptic input from the motor cortex may also be a factor. However, this effect was negligible in the present study, as depressions were not observed in Phase OFF-1 even though rats had already pressed the lever (Fig. 7D). Third, inhibitory presynaptic inputs from the limbic system and/or subcortical areas are other possible mechanisms underlying the slopeTD, and appear most likely to be responsible for it. With this inhibitory system acting as a selective gate allowing important information to pass through while blocking extraneous information from reaching the dentate gyrus, acquisition of operant behavior is facilitated. It may thus be useful to concentrate on learning behavior. Paradoxically, the sensory information reaching the granule cells of the dentate gyrus might be important for the acquisition of operant behavior. Although periodic inhibition was observed in the slope of fEPSPs recorded from the molecular layer, periodic changes were not observed in activation of the slope of fEPSPs and inhibition of the amplitude of pop. spikes recorded from the granule cell layer. We performed each experiment in different animals, and continued to the 30th trial for recording of fEPSPs from the molecular layer and to the 6th trial for recording from the granule cell layer. Six trials might have been insufficient to detect periodicity. Hippocampal LTP is widely believed to contribute to the cellular mechanisms underlying learning and memory, as it has been demonstrated that brief high-frequency stimulation of afferent fibers produces long-lasting enhancement of synaptic efficacy. However, we observed no LTP in this study although our rats were capable of having LTP induced by a high-frequency stimulus. The hippocampal slopeTD induced without any artificial stimulation in this study is thought to be another neural mechanism underlying learning and memory. In conclusion, the present study showed that slopeTD was periodically induced in the molecular layer, slopeTP was induced in the granule cell layer in the dentate gyrus, and consequently spikeTD

was induced in granule cells during the process of acquisition of appetite-driven operant behavior in freely moving rats. SlopeTD might be important for learning and memory. The mechanisms of slopeTD and its periodic appearance remain unknown, and further studies are needed to clarify them. In such studies, the slopes of fEPSPs and the amplitude of pop. spikes should be recorded simultaneously from the molecular layer and the granule cell layer in the same freely moving rat. Acknowledgments This work was supported in part by a grant from the SECOM Science and Technology Foundation, Tokyo. References [1] Bliss TVP, Lømo T. Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J Physiol 1973;232:331–56. [2] Bliss TVP, Gardner-Medwin AR. Long-lasting potentiation of synaptic transmission in the dentate area of the unanaesthetized rabbit following stimulation of the perforant path. J Physiol 1973;232:357–74. [3] Bliss TVP, Collingridge GL. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 1993;361:31–9. [4] Doyère V, Rédini-Del Negro C, Dutrieux G, Le Floch G, Davis S, Laroche S. Potentiation or depression of synaptic efficacy in the dentate gyrus is determined by the relationship between the conditioned and unconditioned stimulus in a classical conditioning paradigm in rats. Behav Brain Res 1995;70:15–29. [5] Weisz DJ, Clark GA, Thompson RF. Increased responsivity of dentate granule cells during nictitating membrane response conditioning in rabbit. Behav Brain Res 1984;12:145–54. [6] Matthies H, Ruethrich H, Ott T, Matthies HK, Matthies R. Low frequency perforant path stimulation as a conditioned stimulus demonstrates correlations between long-term synaptic potentiation and learning. Physiol Behav 1986;36:811–21. [7] Skelton RW, Scarth AS, Wilkie DM, Miller JJ, Phillips AG. Long-term increase in dentate granule cell responsivity accompany operant conditioning. J Neurosci 1987;7:3081–7. [8] Johnston D, Hippocampus Amaral D. In: Shepherd GM, editor. Synaptic organization of the brain. New York: Oxford Univ Press; 2004. p. 455–98. [9] Andersen P, Sundberg SH, Sveen O, Swann JW, Wigström H. Possible mechanism for long-lasting potentiation of synaptic transmission in hippocampal slices from guinea-pigs. J Physiol 1980;302:463–82. [10] Blackstand TW. Commissural connections of the hippocampal region in the rat, with special reference to their mode of termination. J Comp Neurol 1956;105:417–537. [11] Laurberg S, Sorensen KE. Associational and commissural collaterals of neurons in the hippocampal formation (hilus fasciae dentate and subfield CA3). Brain Res 1981;212:287–300. [12] Laatsch RH, Cowan WM. Electron microscopic studies of the dentate gyrus of the rat. II. Degeneration of commissural afferents. J Comp Neurol 1967;130:241–62. [13] Kishi K, Stanfield BB, Cowan WM. A quantitative EM autoradiographic study of the commissural and associational connections of the dentate gyrus in the rat. Anat Embryol 1980;160:173–86. [14] Soriano E, Frotscher M. Spiny nonpyramidal neurons in the CA3 region of the rat hippocampus are glutamate-like immunoreactive and receive convergent mossy fiber input. J Comp Neurol 1993;333:435–48. [15] Kosaka T, Hama K, Wu JY. GABAergic synaptic boutons in the granule cell layer of rata dentate gyrus. Brain Res 1984;293:353–9. [16] Kosaka T. Axon initial segments of the granule cell in the rat dentate gyrus: synaptic contacts on bundles of axon initial segments. Brain Res 1983;274:129–34. [17] Soriano E, Frotscher M. A GABAergic axo-axonic cell in the fascia dentate controls the main excitatory hippocampal pathway. Brain Res 1989;503:170–4. [18] Morrison JH, Benoit R, Magistretti PJ, Ling N, Bloom FE. Immunohistochemical distribution of pro-somatostatin-related peptides in hippocampus. Neurosci Lett 1982;34:137–42. [19] Bakst I, Avendano C, Morrison JH, Amaral DG. An experimental analysis of the origins of somatostatin-like immunoreactivity in the dentate gyrus of the rat. J Neurosci 1986;6:1452–62. [20] Freund TF, Buzsáki G. Interneurons of the hippocampus. Hippocampus 1996;6: 347–470. [21] Amaral DG, Witter MP. Hippocampal formation. In: Paxinos G, editor. The rat nervous system. San Diego: Academic Press; 1995. p. 443–93. [22] Mishkin M, Appenzeller T. The anatomy of memory. Sci Am 1987;256(6):62–71. [23] Hatfield T, Han J-S, Conley M, Gallagher M, Holland P. Neurotoxic lesions of basolateral, but not central, amygdala interfere with Pavlovian second-order conditioning and reinforcer devaluation effects. J Neurosci 1996;16:5256–65. [24] Kesner RP, Martinez Jr JL. Neurobiological views of memory. In: Kesner RP, editor. Learning and memory: a biological view. San Diego: Academic Press; 1991. p. 499–547. [25] Poremba A, Gabriel M. Amygdalar lesions block discriminative avoidance learning and cingulothalamic training-induced neuronal plasticity in rabbits. J Neurosci 1997;17:5237–44.

S. Nomoto et al. / Physiology & Behavior 105 (2012) 938–947 [26] Schoenbaum G, Chiba AA, Gallagher M. Orbitofrontal cortex and basolateral amygdala encode expected outcomes during learning. Nat Neurosci 1998;1:155–9. [27] Akirav I, Richter-Levin G. Biphasic modulation of hippocampal plasticity by behavioral stress and basolateral amygdala stimulation in the rat. J Neurosci 1999;19:10530–5. [28] Carmichael ST, Price JL. Connectional networks within the orbital and medial prefrontal cortex of macaque monkeys. J Comp Neurol 1996;371:179–207. [29] Finch DM. Hippocampal, subicular, and entorhinal afferents and synaptic integration in rodent cingulated cortex. In: Vogt BA, Gabriel M, editors. Neurobiology of cingulate cortex and limbic thalamus. Boston: Birkhäuser; 1993. p. 224–48. [30] Gaffan D, Murray EA. Amygdalar interaction with the mediodorsal nucleus of thalamus and the ventromedial prefrontal cortex in stimulus-reward associative learning in the monkey. J Neurosci 1990;10:3479–93. [31] Groenewegen HJ. Organization of the afferent connections of the mediodorsal thalamic nucleus in the rat, related to the mediodorsal–prefrontal topography. Neuroscience 1988;24:379–431. [32] Saper CB. Organization of cerebral cortical afferent systems in the rat. II. Hypothalamocortical projections. J Comp Neurol 1985;237:21–46. [33] van Groen T, Wyss JM. Extrinsic projections from area CA1 of the rat hippocampus: olfactory, cortical, subcortical, and bilateral hippocampal formation prejections. J Comp Neurol 1990;302:515–28.

947

[34] Zeng D, Stuesse SL. Morphological heterogeneity within the cingulate cortex in rat: a horseradish peroxidase transport study. Brain Res 1991;565:290–300. [35] Papez JW. A proposed mechanism of emotion. 1937. J Neuropsychiatry Clin Neurosci 1995;7:103–12. [36] Vogt BA, Finch DM, Olson CR. Functional heterogeneity in cingulate cortex: the anterior executive and posterior evaluative regions. Cereb Cortex 1992;2:435–43. [37] Devinsky O, Morrell MJ, Vogt BA. Contributions of anterior cingulate cortex to behaviour. Brain 1995;118:279–306. [38] Poremba A, Gabriel M. Medial geniculate lesions block amygdalar and cingulothalamic learning-related neuronal activity. J Neurosci 1997;17: 8645–55. [39] Frysztak RJ, Neafsey EJ. The effect of medial frontal cortex lesions on cardiovascular conditioned emotional responses in the rat. Brain Res 1994;643:181–93. [40] Holson RR. Mesial prefrontal cortical lesions and timidity in rats. I. Reactivity to aversive stimuli. Physiol Behav 1986;37:221–30. [41] Jinks AL, McGregor IS. Modulation of anxiety-related behaviours following lesions of the prelimbic or infralimbic cortex in the rat. Brain Res 1997;772:181–90. [42] Takenouchi K, Nishijo H, Uwano T, Tamura R, Takigawa M, Ono T. Emotional and behavioral correlates of the anterior cingulated cortex during associative learning in rats. Neuroscience 1999;93:1271–87. [43] Iijima T, Witter MP, Ichikawa M, Tominaga T, Kajiwara R, Matsumoto G. Entorhinal– hippocampal interactions revealed by real-time imaging. Science 1996;272:1176–9.