HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 98/4/2196    most recent 00524.2007v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Martin, C. Articles by Kay, L. M. Search for Related Content PubMed PubMed Citation Articles by Martin, C. Articles by Kay, L. M.

An Olfacto-Hippocampal Network Is Dynamically Involved in Odor-Discrimination Learning

¹Department of Psychology and ²Institute for Mind and Biology, The University of Chicago, Chicago, Illinois

Submitted 9 May 2007; accepted in final form 10 August 2007

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Several studies have shown that memory consolidation relies partly on interactions between sensory and limbic areas. The functional loop formed by the olfactory system and the hippocampus represents an experimentally tractable model that can provide insight into this question. It had been shown previously that odor-learning associated beta band oscillations (15–30 Hz) of the local field potential in the rat olfactory system are enhanced with criterion performance, but it was unknown if these involve networks beyond the olfactory system. We recorded local field potentials from the olfactory bulb (OB) and dorsal and ventral hippocampus during acquisition of odor discriminations in a go/no-go task. These regions showed increased beta oscillation power during odor sampling, accompanied by a coherence increase in this frequency band between the OB and both hippocampal subfields. This coherence between the OB and the hippocampus increased with the onset of the first rule transfer to a new odor set and remained high for all learning phases and subsequent odor sets. However, coherence between the two hippocampal fields reset to baseline levels with each new odor set and increased again with criterion performance. These data support hippocampal involvement in the network underlying odor-discrimination learning and also suggest that cooperation between the dorsal and ventral hippocampus varies with learning progress. Oscillatory activity in the beta range may thus provide a mechanism by which these areas are linked during memory consolidation, similar to proposed roles of beta oscillations in other systems with long-range connections.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Contemporary theories propose that long-term memory consolidation is facilitated by temporary bidirectional interactions between sensory cortices and hippocampal regions (Buzsaki 1996; Eichenbaum 2000; Nadel and Moscovitch 1997; Squire and Alvarez 1995; Wiltgen et al. 2004). Because the two systems share strong bidirectional connections, the loop formed by the olfactory system and the hippocampus (HPC) represents a very good model to assess functional connections between the two levels during learning. Anatomically, only two synapses separate the olfactory bulb (OB), the first relay of odor processing, from the dentate gyrus (DG) of the hippocampus via the lateral entorhinal cortex and the lateral perforant path (Vanderwolf 1992). The OB also reaches the entorhinal cortex via the piriform cortex (PC), which has been shown to be involved in storage and retrieval of olfactory memory traces (Barkai and Saar 2001; Haberly 2001). In turn, the HPC sends projections to the OB through the entorhinal cortex (EC) and also directly from the ventral HPC (CA1 area) to the OB granule cell layer (Gulyas et al. 1998; van Groen and Wyss 1990). Connected through strong intrahippocampal connections (Amaral and Witter 1989; Moser and Moser 1998), the dorsal (septal) and ventral (temporal) hippocampal subfields are anatomically distinct and are characterized by functional dissociations, mostly studied in relation to fear conditioning (Richmond et al. 1999; Yoon and Otto 2007).

One mechanism capable of producing functional long-range interactions across such a network is synchronous oscillatory activity as it may create a functional link between two remote neuronal populations or brain areas and thus produce a temporal window for transient communication (Fries 2005). In addition, because large-scale population cooperativity implies coincident neuronal activity arriving at downstream targets, oscillatory synchrony could produce the kind of firing precision necessary to enhance synaptic efficiency leading to network plasticity (Schaefer et al. 2006; Singer 1993). An increasing number of studies in various sensory systems have found a clear relationship between neural population activity and perception (Kay 2003; Liu and Newsome 2006; Mehring et al. 2003; Pesaran et al. 2002) and memory storage (Herrmann et al. 2004; Tallon-Baudry and Bertrand 1999).

Both the OB and HPC show prominent oscillatory modes that share common properties. In the OB, oscillatory activity associated with respiratory drive and afferent input has a frequency range (2–12 Hz) that overlaps with the theta rhythm in the hippocampus (4–12 Hz) (Kay 2005; Macrides et al. 1982). Although in this frequency band hippocampal and OB oscillations are usually uncorrelated, a previous study reported that the olfactory system and the HPC were linked in the high theta frequency band (6–12 Hz) associated with performance accuracy in an olfactory sensorimotor discrimination task (Kay 2005). Sniffing has also been shown to be associated with hippocampal theta rhythm during odor contingency reversal learning in a go/no-go task (Macrides et al. 1982). Other rhythms, such as beta (15–35 Hz) and gamma (40–100 Hz) oscillations have been recorded in both structures and have also been found to be related to behavior (Csicsvari et al. 2003; Kay 2003; Kay and Freeman 1998; Vanderwolf 2001). In the OB and piriform cortex, acquisition of odor discrimination in a go/no-go task leads to the emergence of powerful odor-induced beta oscillations (15–40 Hz) (Martin et al. 2004; Ravel et al. 2003). Intact feedback projections to the OB are necessary for the expression of this activity (Martin et al. 2006), which suggests a long-range cooperative network. Our hypothesis is that beta oscillations link the different elements of a network involved in acquisition of a relation between an odor and a behavioral significance.

In the present study, we describe beta band coupling between the hippocampus and the OB associated with odor-guided associative learning. We recorded the local field potential (LFP) simultaneously in the OB and the dorsal and ventral hippocampus of rats engaged in an olfactory go/no-go task involving appetitive reinforcement. To describe dynamics of functional connectivity between the OB and the HPC, we analyzed characteristics of the oscillatory activity and its interregional coherence during learning and criterion performance of odor pair discriminations. We found an increase of beta band power in both the dorsal and ventral hippocampus during odor sampling, an increase in coherence with the OB specifically during odor sampling, and an increase in intra-hippocampal coherence related to criterion performance.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Four adult male Sprague-Dawley rats (375–425 g; purchased from Harlan, Indianapolis, IN) were implanted with electrodes after a short shaping period (learning to obtain a reward without any odor experience). After recovery they were then trained in a go/no-go task (GNG) to discriminate sets of two odors successively. For the training, animals were dieted to and maintained at 85% of their ad libitum weight. All procedures were carried out with approval from and oversight by the University of Chicago Institutional Animal Care and Use Committee, according to guidelines set by Association for Assessment and Accreditation of Laboratory Animal Care.

Electrophysiological recording

Animals were implanted with stainless steel formvar-coated recording electrodes (100 µm diam, 100–500 k, California Fine Wire; inter-tip vertical distance 0.5–1.0 mm) under pentobarbital anesthesia. Electrodes were positioned stereotaxically in the lateral olfactory tract (LOT) (2.7 mm anterior to bregma, 1.6 mm lateral from midline, 14 deg from vertical, depth 8 mm), OB (8.5 mm anterior to bregma, 1.5 mm lateral, depth 4 mm), the dorsal part of the HPC at the level of the DG (dHPC; 4.8 mm posterior to bregma, 2 mm lateral, depth 3 mm), and in the ventral part of the HPC (vHPC; 5.8 mm posterior to bregma, 2 mm lateral, 20° from vertical, depth 9 mm). The depth of placements was positioned at the level of the principal cell layers in the OB and HPC using both multiunit activity and the evoked field potential induced in response to electrical stimulation of the LOT (10 V; 0.02- to 0.08-ms duration; 0.5-Hz stimulation rate).

The reference and ground electrodes were connected to skull screws located above the posterior portion of the contralateral cortical hemisphere. Electrodes were inserted into a threaded round nine-hole connector (Ginder Scientific) and fixed onto the rat's head with dental acrylic. Two weeks of recovery separated surgery from recording sessions.

For each rat, electrophysiological measures were recorded every session throughout training (except for presurgical initial shaping). Neural data from the different electrodes along with behavioral event markers were recorded with a Neuralynx Cheetah 32 channel system. Signals were sampled at 2,016 Hz, amplified (x2,000), and analog filters were set at 1–475 Hz. A unity gain preamplifier headstage (NB Labs, Denison, TX) was used for signal conditioning.

Verification of electrode placements

After experiments were complete, rats were injected with a lethal dose of urethane, and electrode placements were marked by passing current through the tip of the stainless steel wires. Electrode tips were marked using the Prussian Blue reaction, as rats were perfused intracardially with a 10% formalin solution with 4% potassium ferrocyanide. Brains were extracted from the skull, sectioned coronally, and stained with Neutral Red. As illustrated in Fig. 1, recording electrode placements were at the level of the mitral cell layer in the OB, at the vicinity of the hilus in the dorsal HPC, and in the CA1 layer in the ventral HPC. However, in the dHPC, considering the relatively large distance covered by LFP recording we cannot exclude some influence from CA1.

View larger version (40K): [in this window] [in a new window]   FIG. 1. Electrode placement (dots represent localization of the tips for the 4 animals) reported on schematic brain sections (adapted from Paxinos and Watson 1997).

Experiments were conducted in a 30 x 25 x 50-cm operant chamber (MedAssociates, St. Albans, VT) within a Faraday cage. The front wall of the chamber contained an odor port, under which protruded a pellet delivery tube. A retractable lever was located on the right side of the port. A light was positioned on the top of the back wall. Behavioral events and responses of the animal were controlled and monitored by a separate computer (Coulbourn Graphic State Allentown, PA), and event markers were transmitted directly to the Neuralynx recording system via the digital IO line. Odors were generated in glass test tubes by bubbling air (100 ml/min) through a column of pure odorant solutions and injecting the odorized air into a carrier air stream (400 ml/min) via a computer-controlled olfactometer. For all odorants, odorized air was diluted to 20% of saturated vapor with a continuous plain air stream. Because the odorants were of various volatilities, producing saturated vapor concentrations related to the vapor pressure, the absolute airborne concentrations of odorants varied.

Illumination of the light indicated that the odor port was active. The rat initiated the odor delivery, which was triggered by detection of a nose poke in the odor port. One of two possible odors, assigned to valence CS+ or CS–, was then pseudorandomly delivered for 1.5 s. One second after the end of the odor pulse, the lever extended from the wall. If the odor was CS+, pressing the lever delivered a sucrose pellet (45 mg; Research Diets and Bioserve). Pressing the lever for odor CS– switched off the light and doubled the intertrial interval from 7 to 14 s. The rats learned to press the lever (go response) following sampling of odor CS+ and to avoid pressing the lever (no-go response) following sampling of odor CS– (Fig. 2, A and B).

View larger version (20K): [in this window] [in a new window]   FIG. 2. Behavioral protocol and results. A: behavioral paradigm. Schematic picturing the time course of a single trial (see description in METHODS). B: example of a latency curve for 1 rat, during the acquisition of geraniol(+)/citral(–) (blocks 1–10; each block is 20 trials) and the reversal (block 11 to end). Each value represents the average of blocks of 10 trials of the latency between odor onset and lever press measured for every trial (correct no-go is assigned the maximum latency of 8.8 s). C: mean number of trials (±SE) performed by the animals to reach the criterion level for each odor pair, in chronological order from left to right. Transfers (pairs 2 and 3) took significantly fewer trials than the 1st pair; in contrast, reversal took significantly longer.

Odors used were octanol-propanol (Fisher Scientific, Fair Lawn, NJ; pair 1), hexanol-heptanol (ICN Biomedicals, Aurora, OH; pair 2), geraniol (Fisher Scientific, Fair Lawn NJ)-citral (Fluka Chemika; pair 3). After pair 3 was completed, we reversed the valence attached to the odors, so that geraniol CS+ became CS– and inversely for citral (reversal).

Data analysis

LFP signals were analyzed off-line with IGOR Pro 5.03 (Wavemetrics, Portland, OR). Analysis was focused on a 4-s interval surrounding the odor onset. Data were inspected to discard trials containing movement artifacts (5% of trials); these are easily recognizable by a simple visual examination of raw signals. Spectral analysis was done on the raw analog filtered 1- to 475-Hz signal.

POWER ANALYSIS.  Auto-spectra were estimated by first applying a 1,024-point Hamming taper to each data window, then taking the fast Fourier transform (FFT) and computing the spectrum. Half-overlapping 1,024 point windows were used to obtain the averaged power spectrum for the 2-s preodor and odor period time segments. Total power was obtained by integrating the band between 15 and 35 Hz. Power spectra were averaged for the two periods across the 20 (beginner level) or 40 (criterion and postcriterion levels) consecutive trials corresponding to a given performance level. Dynamic power spectra centered on odor delivery were estimated for each 1,024 point window stepped by 250 ms.

COHERENCE ANALYSIS.  Coherence spectra between the pairs of sites were calculated as the normalized cross-spectral density of two waves using a FFT window of 1,024 samples as has been reported elsewhere (Kay and Freeman 1998; Lowry and Kay 2007). Because coherence is the cross-spectrum normalized by the autospectra of the two signals, it is not affected by the absolute power of these signals. Coherence is then a measure for the interdependence of two signals in the frequency domain with values between zero and one. A value of zero means that the two signals are independent at the considered frequency, whereas a value of 1 means that the signals are identical in frequency and have a constant phase relationship. Enhancement of coherence signifies an increase of frequency similarity and phase consistency between oscillatory signals in two brain regions.

Statistical analyses for beta frequency (15–35 Hz) coherence were done with a FFT window of 1,024 samples. Averaged coherence estimates were obtained by averaging the values in the band between 15 and 35 Hz. Then the values were averaged for the preodor and the odor period across the 20 (beginner level) or 40 (criterion and postcriterion levels) consecutive trials for a given odor-contingency pair corresponding to the given level.

Statistical analyses

For behavioral results, we assessed whether the time necessary for the animals to learn the discrimination was different according to odor pair. The number of trials necessary to reach criterion was compared across pairs using a t-test.

Statistical analysis of power and coherence changes were done by two different methods. As the distribution of power values is not normal, statistical comparisons were performed with nonparametric tests. First the difference between values of preodor and odor time windows was tested for each reward condition, training level, and odor using a Wicoxon signed-rank test. Then the difference between beginner versus criterion and criterion versus postcriterion were assessed separately using the power ratio (odor period/preodor period). All pairwise differences were tested using the Mann–Whitney U test for unmatched samples.

Coherence values were normally distributed (Kolmogoroff-Smirnoff Normality test, P > 0.9), so repeated-measures ANOVAs were performed to assess significant differences during the odor sampling period versus the preodor period (significance level set at P < 0.05). Bonferroni post hoc tests were used to assess pairwise comparisons. Two independent factors were tested: the level of training (beginner, criterion, postcriterion) and the odor-contingency value (CS+ and CS–).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
LFPs were recorded in the OB (ventral portion near the mitral cell layer), dorsal HPC (dHPC, near the hilus of the DG), and ventral HPC (vHPC; in CA1; Fig. 1) while rats performed a two odor GNG discrimination task. After the task was learned with a first odor pair, it was then transferred successively to two other pairs, and the final pair was then reversed in contingency.

Acquisition of the first pair of odors took significantly longer than the others, with an average of 348 ± 72 trials for pair 1 [octanol(+)/propanol(–)], compared with 72 ± 41 and 128 ± 84 trials for pair 2 [hexanol(+)/heptanol(–)] and pair 3 [geraniol(+)/citral(–)], respectively [pairs 1 and 2, t(5) = –6.472 P < 0.005; pairs 1–3, t(5) = –3.625 P < 0.05; pairs 2 and 3, t(6) = 1.21 P > 0.1]. However, the number of trials to reach criterion for the reversal of pair 3 (423 ± 53 trials) was not different from acquisition of pair 1 [t(5) = 1.619 P > 0.1, geraniol(+)/citral(–) vs. citral(+)/geraniol(–) t(6) = –5.980 P = 0.001; Fig. 2, B and C].

To compare changes in the LFP signal associated with odor-discrimination learning, we defined three behavioral levels for every odor-contingency pair: beginner, the first 20 presentations of each odor in the pair (CS+ and CS–); criterion, the 40 trials of each odor starting with the first 90% correct no-go block; and postcriterion, the 40 trials of each odor just after the criterion level.

Modifications of beta band activity during odor sampling

As shown in Fig. 3, during spontaneous activity outside of the odor sampling periods, we observed typical OB LFP activity, i.e., bursts of gamma activity (60–100 Hz) superimposed on slow waves related to respiration. In the HPC, the power spectrum was dominated by the theta frequency (4–12 Hz; Fig. 3, A and B). As soon as the behavioral performance of the animals improved for each odor set, beta oscillations were evident in the OB raw signal during odor sampling (Fig. 3A). Hippocampal signals also contained observable beta oscillatory activity in the same time period with smaller amplitudes than in the OB (Fig. 3A).

View larger version (41K): [in this window] [in a new window]   FIG. 3. Raw traces and dynamic power spectra shows an odor induced increase in beta (15–35 Hz) power in olfactory bulb (OB), dorsal and ventral parts of the hippocampus at the level of the dentate gyrus (dHPC and vHPC) accompanied by an increase in coherence between these regions in the same frequency band. A: raw local field potential (LFP) signal (1–475 Hz) recorded in the OB (top), dHPC (middle), and vHPC (bottom) from single trials, during sampling of heptanol(+) at the criterion level. In the absence of odor, lower amplitude gamma bursts (60–90 Hz) are present at each respiratory cycle in the OB (inhalations are downward deflections in the signal, arrowheads point at bursts of gamma activity present at each respiratory cycle). The vertical blue line marks the odor onset. An increase in the beta oscillatory response can be observed during odor sampling. B: dynamic power spectra show elevation of beta power (left) and coherence (right) during odor sampling for each of the recording locations and pairs in the beta frequency band. Theta band activity is evident in the spectra from the hippocampal regions. Shown are averages across the criterion level (40 trials) of trials from the same rat for sampling of heptanol(+). Odor onset is indicated by the vertical white line at 0 s. x axis, time (in s); y axis, frequency from 0 to 60 Hz from top to bottom. Left: color scales represent signal power (µV2); right: scale represents coherence.

View larger version (36K): [in this window] [in a new window]   FIG. 4. Power spectra modifications between preodor and odor period. Top: OB power spectra averaged across all rats over criterion trials (40 trials/rat/odor) during the odor delivery period for CS+ (thick line) and CS– (dashed line), and for the preodor period (CS+ and CS– trials pooled, thin dotted line). The region covering the power in the beta frequency band is expanded in the cutouts in each plot. Middle and bottom: dHPC and vHPC power spectra averaged across all rats over criterion trials (40 trials/rat/odor) pooled for CS+ and CS– during odor delivery period (thick line) and preodor period (thin line). Spectra were calculated using half-overlapping fast Fourier transform (FFT) windows of 1,024 samples to cover the 2-s periods. Vertical bars indicate ±1 SE.

To analyze beta band (15–35 Hz) activity elicited by odor sampling, we compared the 2-s odor period starting with the odor onset (odor period) with the 2 s just prior to odor onset (preodor period). Analysis performed animal by animal showed that for all behavioral levels (beginner, criterion, postcriterion), during odor sampling there was a significant power increase in the beta frequency band relative to the preodor period (Wilcoxon signed-rank test P < 0.001), except for one animal in criterion and postcriterion levels for the first odor pair (octanol-propanol P = 0.20).

To compare beta power values across odors, behavioral levels, and rats, we assessed the ratio of power between the odor and preodor periods. In the OB, odor-elicited beta power showed a strong correlation with learning and, aside from variations in amplitude with different odor sets, showed an increase from beginner to criterion levels (Fig. 5; Mann-Whitney U test, pair 1: P < 0.001; pair 2: P < 0.001 for CS+, n.s. for CS–; pair 3: P < 0.001; reversal: P < 0.001 for CS+ and CS–). Beta power did not change from beginner to criterion levels for pair 2 CS- (hexanol), as beta power was already high in the beginner level, due to a very early increase in power for this odor (in the 1st 5 trials) followed by a very quick onset of >70% correct responding. In addition, except for odor pair 3 [geraniol(+)/citral(–)], there was also a decrease in beta power from criterion to postcriterion level (Fig. 3; Mann-Whitney U test, pair 1: P = 0.019; pair 2: P < 0.001; pair 3: n.s.; reversal: P < 0.001).

View larger version (14K): [in this window] [in a new window]   FIG. 5. Beta power increase during odor sampling in the OB is learning related. Power extracted for odor sampling (thick lines represent the CS+ odor, dashed lines the CS– odor) and preodor period (CS+ and CS– trials pooled, thin line) in the beta frequency band (15–35 Hz) shows that activity in this frequency range increases from beginner (beg) to criterion (crit) and tend to decreases after criterion (post). Values for each behavioral level represent the average across all animals. The increase of power between the preodor and odor periods is significant for all conditions. Triangle, significant difference between the 2 odors of the pair (P < 0.05); pound sign, significant differences between the beginner and criterion levels or between the criterion and postcriterion levels; plus sign, that the difference between beginner and criterion is significant only for heptanol in the odor set (pair 2).

Beta power increases during odor sampling in the HPC

During odor sampling, the HPC also showed an increase in beta band power although it did not span as large a range as in the OB (Figs. 4 and 6). For all the animals and odor sets, both in the dorsal and ventral HPC, beta power was significantly higher for the odor period compared with preodor period (Wilcoxon signed-rank test P < 0.001). Figure 6 summarizes analyses conducted separately for every behavioral level. It shows that this increase was significant (P < 0.01) for most of the conditions. The increase was not significant for beginner and criterion levels for pair 3 [geraniol(+)-citral(–)], and for the same odors just prior to reversal [citral(+)-geraniol(–)].

View larger version (21K): [in this window] [in a new window]   FIG. 6. In the HPC, beta power (15–35 Hz) increased between the preodor and odor periods but showed no reproducible relation with learning levels (beginners, beg; criterion, crit; and postcriterion, post). Values represent averages across all animals, thick lines represent the CS+ odor, dashed lines the CS– odor, and thin line the preodor period (CS+ and CS– trials pooled). Asterisk, significant difference between preodor and odor period (P < 0.01); plus sign, indicates that the difference between preodor and odor period is significant only for CS+. Significant learning level differences are not consistent and thus are not shown on the figure.

Coherence increases between OB and HPC subfields in the beta band during odor sampling

Coherence measures allow us to estimate changes in cooperativity between brain areas independent of amplitude or power differences. Figure 7 (top and middle) displays the coherence values between OB and HPC subfields from the odor and the preodor periods during criterion level averaged over all rats. This confirms the specificity of the beta frequency range for a coherence increase.

View larger version (38K): [in this window] [in a new window]   FIG. 7. Modifications in coherence spectra among OB-dHPC, OB-vHPC, and dHPC-vHPC between preodor and odor periods are restricted to the beta frequency range (15–40 Hz). Coherence spectra averaged across all rats over criterion trials (40 trials/rat/odor) for the odor delivery period for CS+ (thick line) and CS– (dashed line) odors, and for the preodor period (CS+ and CS– trials pooled, thin dotted line). Spectra were calculated using FFT windows of 1,024 samples. Vertical bars indicate ±1 SE. Coherence spectra between OB and dHPC (top), OB and vHPC (middle), and dHPC and vHPC (bottom) are presented for octanol-propanol, heptanol-hexanol, geraniol-citral, and the reversal (citral-geraniol) from left to right. Except for pair 1, the spectra reveal an increase of coherence for the odor period compared with the preodor period in the beta frequency band (15–35 Hz).

View larger version (24K): [in this window] [in a new window]   FIG. 8. Odor-induced beta band (15–35 Hz) coherence between OB-dHPC (top), OB-vHPC (middle), and dHPC-vHPC (bottom) for the different odor sets (from left to right). An odor induced beta coherence increase is not observed for the 1st odor set. Shown are coherence values displayed for the odor and preodor periods. The increase of coherence during the odor sampling period (thick solid lines represent CS+ odor; dashed lines, CS– odor) compared with the preodor period (thin line represents values pooled for CS+ and CS– trials) for beginner and criterion learning levels. Values are averages across all rats. Asterisk, significant difference between the preodor and odor periods (P < 0.001); triangle, significant difference between CS+ and CS– odor; and pound sign, significant difference between beginner and criterion learning levels (P < 0.05).

Coherence between dHPC and vHPC

Coherence between the two hippocampal subfields showed differences from the OB-HPC patterns. As expected and as can be seen in Fig. 7 (bottom), coherence values in the preodor period were already higher than OB-HPC pairs. Also notice the very high values in theta band coherence for both preodor and odor periods.

Focusing on the beta band (Fig. 8, bottom) one notable feature is that for all four odor-contingency pairs, coherence between the two hippocampal subfields increased from the beginner level to the criterion level [with a significant interaction with preodor/odor period increase except for pair 1: pair 1, F_level(1,476) = 9.305 P < 0.005; pair 2, F_{pre/odor*level}(1,396) = 5.968 P < 0.05; pair 3, F_{pre/odor*level}(1,476) = 4.183 P < 0.05; reversal, F_{pre/odor*level}(1,476) = 3.773 P = 0.05]. Thus we conducted separate analyses for beginner and expert levels to assess the coherence increase during odor sampling relative to the preodor period. Indeed, the odor-evoked coherence increase was never significant in the beginner level (P > 0.1). Beginning with the second learned pair at criterion, odor sampling produced an increase in beta band coherence compared with the preodor period [pair 2, F_pre/odor(1,318) = 55.921, P < 0.001; pair 3, F_pre/odor(1,318) = 13.782, P < 0.0005; reversal, F_pre/odor(1,318) = 36.375, P < 0.0001, no interaction with odor].

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Our results suggest a functional coupling between the OB and the dorsal and ventral parts of the HPC, related to odor-discrimination learning. LFPs were recorded in these structures in awake behaving rats learning to perform successive odor discriminations in a GNG task. We confirm the close relation between beta activity in the OB, the first relay of olfactory processing, and learning (Figs. 3 Figs. 4 Figs. 5) (Martin et al. 2004; Ravel et al. 2003), and for the first time, we show an increase in beta band (15–35 Hz) oscillatory power in the dorsal and ventral HPC and an increase in coherence between the OB and HPC and between the two subfields of the HPC in the same frequency band during the learning task. All of these effects occur specifically during odor sampling.

Beta oscillation power in the OB and HPC

Beta frequency (15–35 Hz) activity has previously been observed in the olfacto-limbic circuit during presentations of behaviorally significant odors (Zibrowski and Vanderwolf 1997), following repeated presentations of the same odor (Gray and Skinner 1988; Lowry and Kay 2007; Vanderwolf and Zibrowski 2001) or for odors experimentally associated with a reward (Martin et al. 2004). In the present report, for every new odor-contingency pair, the same transient increase in beta power during odor sampling occurred in the OB with different magnitudes for different odor sets (Fig. 5). That the OB beta power increase occurs during the build-up of every new association and diminishes once this representation is formed confirms that these oscillations are related to associative odor-discrimination learning in the GNG task and are not just part of a simple familiarization or sensitization process or due to vapor phase concentration (Lowry and Kay 2007).

The involvement of the HPC was different from that of the OB. Beta oscillation power within HPC subfields did not show significant changes with respect to learning level, as beta band power was stable from beginner to criterion levels for most of the odors (Fig. 6). This discounts the hypothesis that the beta increase in the HPC is the result of drive or conduction from the OB. We infer that modulation of HPC beta oscillation power may be driven by the task context or by odor sampling rather than by the learning process itself.

Beta band coherence between the OB and HPC

One of the most important results of the present study concerns the beta band (15–35 Hz) coherence increase between the OB and the two hippocampal subfields that occurs precisely during odor sampling (Figs. 7 and 8). As suggested by computational models of the HPC (Kopell et al. 2000; Schaefer et al. 2006; Traub et al. 2004), slower frequencies (in the beta frequency range or less) are more efficient than faster frequencies for inducing plasticity related phenomena, and large-scale oscillatory coupling is an advantage in binding together widespread areas of the brain (Fries 2005). Experimental data have confirmed these effects, in cats (Roelfsema et al. 1997) and humans (Tallon-Baudry et al. 2001), as slow activity is related to long-range synchronization. Thus under the assumption that beta oscillations serve to functionally connect widely separated brain regions, a coherence increase signifies that the HPC is involved in olfactory processing in some conditions.

Hippocampal involvement in the network underlying olfactory learning has been suggested by studies showing odor-specific responses of neurons in the dorsal HPC (Wiebe and Staubli 1999; Wood et al. 1999). Its implication in olfaction was also assessed by Chaillan et al. (1999), who showed that olfactory learning was accompanied by potentiation of a polysynaptic response to electrical stimulation of the LOT in the DG of behaving rats.

In this study, contrary to beta power evolution in the OB, we did not see a reliable and repeated increase in OB-HPC coherence that followed the learning process. Also, for the first odor set, there was not a significant increase in OB-HPC beta band coherence (Figs. 7 and 8). However, early in the first transfer to a new odor set coherence between the OB and the two HPC subfields increased and remained high, without resetting at the beginning of each new odor set (Fig. 8), although within the HPC circuit, coherence increased from beginner to criterion levels (see following text). We cannot exclude the possibility that the lack of coherence in the first odor set is dependent on the odors used. However, beta power is significantly increased during odor sampling in the first set in both the OB and HPC without a related increase in coherence, which argues against this interpretation. These results may be explained by the observation that the HPC is more involved in the transfer of a rule to an olfactory stimulus set than in performing an odor discrimination per se. This is consistent with lesion studies (Bunsey and Eichenbaum 1996). One common property of hippocampal-dependent tasks is the requirement of learning "flexibility," to employ knowledge obtained in one set of circumstances to solve new problems (Cohen and Eichenbaum 1991; Eichenbaum 2001; Eichenbaum et al. 1999; Squire et al. 1993). In our study, for the first odor pair, the rules of the GNG task are learned at the same time as the odor discrimination itself, and acquisition of the following sets requires rule transfer, which very likely involves different networks. Our hypothesis is that a functional link between sensory areas and the HPC is required for the correct behavioral utilization of sensory cues in the specific context where flexibility is needed but not for the sensory process itself.

This hypothesis is supported by other studies. First, it has been shown that hippocampal lesions do not impair rats’ ability to perform an odor-discrimination task, but they cannot express flexible forms of memory, like transitivity (Bunsey and Eichenbaum 1996). Another argument comes from studies of olfactory learning-related cellular modifications (Hess et al. 1995; Knafo et al. 2005), which are detected in the HPC only after acquisition of rule learning. In particular, upregulation of the L1 cell adhesion molecule implicated in synaptic plasticity has been shown to be predictive of rule learning in the piriform cortex but occurs only after completion of rule learning in the HPC (Knafo et al. 2005). That may also explain why no consistent OB-HPC coherence increase was found in the beta band in previous studies (Kay 2005; Kay and Freeman 1998). In those studies, a single odor pair was used; this did not require a rule transfer.

In our analysis of the HPC theta rhythm and oscillations associated with respiration in the same frequency band in the OB, we did not find that coherence magnitude was either positively correlated with performance of a known odor discrimination (Kay 2005) or negatively correlated with performance in contingency reversal learning (Macrides et al. 1982). There were important differences between the previous studies showing positive results and our study. Kay's paper (2005) involved a different version of a GNG task that required the rats to keep track of the timing of the intertrial interval, and the difference between those results and ours may rely on this task demand. There were also differences in the nature of the data analyzed in the study by Macrides et al. (1982). Their analysis focused only on reversal of the first odor pair, whereas ours was for the third odor pair (after rule learning), and instead of the coherence measure they analyzed the phase between nasal airflow and LFPs in the HPC. Although this is similar to a coherence measure, the statistical evaluation was less detailed in that study.

Dorsal and ventral hippocampal involvement in odor sampling and learning

We found that the two hippocampal regions do not always function as a unit during odor learning because we do not observe a constant high coherence between them. Within the HPC there is an effect of learning level on beta band coherence between the dorsal and ventral subfields (Fig. 8). We show that for every odor set, a coherence increase between the dorsal and ventral HPC during odor sampling rose to significance only when the rats reached criterion performance. Our results suggest that as perceptual learning progresses the HPC does function as a unit, but in the early stages for each odor set, the two subfields are functionally less cooperative and may have separate functions.

The dorsal (septal) and ventral (temporal) subfields of the HPC are also anatomically segregated, and the ventral HPC includes a pool of cells (stratum radiatum giant cells) projecting directly to the granule cell layer of the OB (Gulyas et al. 1998; van Groen and Wyss 1990). Functional specificities of the two HPC subfields are not clearly known; lesion studies lead to contradictory results concerning the involvement of the ventral HPC in spatial learning (de Hoz et al. 2003; Moser et al. 1995), and the two subfields may have different contributions in the context of fear conditioning (Yoon and Otto 2007). In our task, the stronger coherence of the OB with the ventral HPC suggests a stronger link between these two regions than between the OB and the dorsal HPC.

To summarize, we propose that the different types of learning (simple odor discrimination vs. transfer and reversal) involve different brain networks and that beta oscillatory activity may support long range communication in the distributed olfacto-hippocampal system. Under some conditions, the odor- and learning-related beta activity characteristic of the olfactory system becomes coherent with the beta activity elicited in the HPC when an odor is sampled, leading to efficient stimulus processing and an adapted behavioral output in this context. Future studies will examine the local circuitry involved in producing these oscillations within the HPC and in interaction with olfactory areas.

	GRANTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Support for this work was from the Fyssen Foundation (Paris) to C. Martin and National Institute of Deafness and Other Communications Disorders Grant R01-DC007995 Collaborative Research in Computational Neuroscience program to L. M. Kay.

	FOOTNOTES

 
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Address for reprint requests and other correspondence: L. M. Kay, Inst. for Mind and Biology, 940 E. 57th St., Chicago, IL 60637 (E-mail: LKay{at}uchicago.edu)

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Amaral DG, Witter MP. The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience 31: 571–591, 1989.[CrossRef][Web of Science][Medline]

Barkai E, Saar D. Cellular correlates of olfactory learning in the rat piriform cortex. Rev Neurosci 12: 111–120, 2001.[Web of Science][Medline]

Bunsey M, Eichenbaum H. Conservation of hippocampal memory function in rats and humans. Nature 379: 255–257, 1996.[CrossRef][Medline]

Buzsaki G. The hippocampo-neocortical dialogue. Cereb Cortex 6: 81–92, 1996.[Abstract/Free Full Text]

Chaillan FA, Truchet B, Roman FS, Soumireu-Mourat B. Early polysynaptic potentiation recorded in the dentate gyrus during an associative learning task. Neuroscience 94: 443–451, 1999.[CrossRef][Web of Science][Medline]

Cohen NJ, Eichenbaum H. The theory that wouldn't die: a critical look at the spatial mapping theory of hippocampal function. Hippocampus 1: 265–268, 1991.[CrossRef][Medline]

Csicsvari J, Jamieson B, Wise KD, Buzsaki G. Mechanisms of gamma oscillations in the hippocampus of the behaving rat. Neuron 37: 311–322, 2003.[CrossRef][Web of Science][Medline]

de Hoz L, Knox J, Morris RG. Longitudinal axis of the hippocampus: both septal and temporal poles of the hippocampus support water maze spatial learning depending on the training protocol. Hippocampus 13: 587–603, 2003.[CrossRef][Web of Science][Medline]

Eichenbaum H. A cortical-hippocampal system for declarative memory. Nat Rev Neurosci 1: 41–50, 2000.[CrossRef][Web of Science][Medline]

Eichenbaum H. The hippocampus and declarative memory: cognitive mechanisms and neural codes. Behav Brain Res 127: 199–207, 2001.[CrossRef][Web of Science][Medline]

Eichenbaum H, Dudchenko P, Wood E, Shapiro M, Tanila H. The hippocampus, memory, and place cells: is it spatial memory or a memory space? Neuron 23: 209–226, 1999.[CrossRef][Web of Science][Medline]

Fries P. A mechanism for cognitive dynamics: neuronal communication through neuronal coherence. Trends Cogn Sci 9: 474–480, 2005.[CrossRef][Web of Science][Medline]

Gray CM, Skinner JE. Field potential response changes in the rabbit olfactory bulb accompany behavioral habituation during the repeated presentation of unreinforced odors. Exp Brain Res 73: 189–197, 1988.[CrossRef][Web of Science][Medline]

Gulyas AI, Toth K, McBain CJ, Freund TF. Stratum radiatum giant cells: a type of principal cell in the rat hippocampus. Eur J Neurosci 10: 3813–3822, 1998.[CrossRef][Web of Science][Medline]

Haberly LB. Parallel-distributed processing in olfactory cortex: new insights from morphological and physiological analysis of neuronal circuitry. Chem Senses 26: 551–576, 2001.[Abstract/Free Full Text]

Herrmann CS, Munk MH, Engel AK. Cognitive functions of gamma-band activity: memory match and utilization. Trends Cogn Sci 8: 347–355, 2004.[CrossRef][Web of Science][Medline]

Hess US, Lynch G, Gall CM. Regional patterns of c-fos mRNA expression in rat hippocampus following exploration of a novel environment versus performance of a well-learned discrimination. J Neurosci 15: 7796–7809, 1995.[Abstract]

Kay LM. Two species of gamma oscillations in the olfactory bulb: dependence on behavioral state and synaptic interactions. J Integrative Neuroscience 2: 31–44, 2003.[CrossRef]

Kay LM. Theta oscillations and sensorimotor performance. Proc Natl Acad Sci USA 102: 3863–3868, 2005.[Abstract/Free Full Text]

Kay LM, Freeman WJ. Bidirectional processing in the olfactory-limbic axis during olfactory behavior. Behav Neurosci 112: 541–553, 1998.[CrossRef][Web of Science][Medline]

Kepecs A, Uchida N, Mainen ZF. The sniff as a unit of olfactory processing. Chem Senses 31: 167–179, 2006.[Abstract/Free Full Text]

Knafo S, Barkai E, Libersat F, Sandi C, Venero C. Dynamics of olfactory learning-induced up-regulation of L1 in the piriform cortex and hippocampus. Eur J Neurosci 21: 581–586, 2005.[CrossRef][Web of Science][Medline]

Kopell N, Ermentrout GB, Whittington MA, Traub RD. Gamma rhythms and beta rhythms have different synchronization properties. Proc Natl Acad Sci USA 97: 1867–1872, 2000.[Abstract/Free Full Text]

Liu J, Newsome WT. Local field potential in cortical area MT: stimulus tuning and behavioral correlations. J Neurosci 26: 7779–7790, 2006.[Abstract/Free Full Text]

Lowry CA, Kay LM. Chemical factors determine olfactory system beta oscillations in waking rats. J Neurophysiol 98: 394–404, 2007.[Abstract/Free Full Text]

Macrides F, Eichenbaum HB, Forbes WB. Temporal relationship between sniffing and the limbic theta rhythm during odor discrimination reversal learning. J Neurosci 2: 1705–1717, 1982.[Web of Science][Medline]

Martin C, Gervais R, Hugues E, Messaoudi B, Ravel N. Learning modulation of odor-induced oscillatory responses in the rat olfactory bulb: a correlate of odor recognition? J Neurosci 24: 389–397, 2004.[Abstract/Free Full Text]

Martin C, Gervais R, Messaoudi B, Ravel N. Learning-induced oscillatory activities correlated to odor recognition: a network activity. Eur J Neurosci 23: 1801–1810, 2006.[CrossRef][Web of Science][Medline]

Mehring C, Rickert J, Vaadia E, Cardosa de Oliveira S, Aertsen A, Rotter S. Inference of hand movements from local field potentials in monkey motor cortex. Nat Neurosci 6: 1253–1254, 2003.[CrossRef][Web of Science][Medline]

Moser MB, Moser EI. Functional differentiation in the hippocampus. Hippocampus 8: 608–619, 1998.[CrossRef][Web of Science][Medline]

Moser MB, Moser EI, Forrest E, Andersen P, Morris RG. Spatial learning with a minislab in the dorsal hippocampus. Proc Natl Acad Sci USA 92: 9697–9701, 1995.[Abstract/Free Full Text]

Nadel L, Moscovitch M. Memory consolidation, retrograde amnesia and the hippocampal complex. Curr Opin Neurobiol 7: 217–227, 1997.[CrossRef][Web of Science][Medline]

Paxinos G, Watson C. The Rat Brain in Stereotaxic Coordinates (3^rd ed.). New York: Academic, 1997.

Pesaran B, Pezaris JS, Sahani M, Mitra PP, Andersen RA. Temporal structure in neuronal activity during working memory in macaque parietal cortex. Nat Neurosci 5: 805–811, 2002.[CrossRef][Web of Science][Medline]

Ravel N, Chabaud P, Martin C, Gaveau V, Hugues E, Tallon-Baudry C, Bertrand O, Gervais R. Olfactory learning modifies the expression of odour-induced oscillatory responses in the gamma (60–90 Hz) and beta (15–40 Hz) bands in the rat olfactory bulb. Eur J Neurosci 17: 350–358, 2003.[CrossRef][Web of Science][Medline]

Richmond MA, Yee BK, Pouzet B, Veenman L, Rawlins JN, Feldon J, Bannerman DM. Dissociating context and space within the hippocampus: effects of complete, dorsal, and ventral excitotoxic hippocampal lesions on conditioned freezing and spatial learning. Behav Neurosci 113: 1189–1203, 1999.[CrossRef][Web of Science][Medline]

Roelfsema PR, Engel AK, Konig P, Singer W. Visuomotor integration is associated with zero time-lag synchronization among cortical areas. Nature 385: 157–161, 1997.[CrossRef][Medline]

Schaefer AT, Angelo K, Spors H, Margrie TW. Neuronal oscillations enhance stimulus discrimination by ensuring action potential precision. PLoS Biol 4: e163, 2006.[CrossRef][Medline]

Singer W. Synchronization of cortical activity and its putative role in information processing and learning. Annu Rev Physiol 55: 349–374, 1993.[CrossRef][Web of Science][Medline]

Squire LR, Alvarez P. Retrograde amnesia and memory consolidation: a neurobiological perspective. Curr Opin Neurobiol 5: 169–177, 1995.[CrossRef][Web of Science][Medline]

Squire LR, Knowlton B, Musen G. The structure and organization of memory. Annu Rev Psychol 44: 453–495, 1993.[CrossRef][Web of Science][Medline]

Tallon-Baudry C, Bertrand O. Oscillatory gamma activity in humans and its role in object representation. Trends Cogn Sci 3: 151–162, 1999.[CrossRef][Web of Science][Medline]

Tallon-Baudry C, Bertrand O, Fischer C. Oscillatory synchrony between human extrastriate areas during visual short-term memory maintenance. J Neurosci 21: RC177, 2001.[Abstract/Free Full Text]

Traub RD, Bibbig A, LeBeau FE, Buhl EH, Whittington MA. Cellular mechanisms of neuronal population oscillations in the hippocampus in vitro. Annu Rev Neurosci 27: 247–278, 2004.[CrossRef][Web of Science][Medline]

van Groen T, Wyss JM. Extrinsic projections from area CA1 of the rat hippocampus: olfactory, cortical, subcortical, and bilateral hippocampal formation projections. J Comp Neurol 302: 515–528, 1990.[CrossRef][Web of Science][Medline]

Vanderwolf CH. Hippocampal activity, olfaction, and sniffing: an olfactory input to the dentate gyrus. Brain Res 593: 197–208, 1992.[CrossRef][Web of Science][Medline]

Vanderwolf CH. The hippocampus as an olfacto-motor mechanism: were the classical anatomists right after all? Behav Brain Res 127: 25–47, 2001.[CrossRef][Web of Science][Medline]

Vanderwolf CH, Zibrowski EM. Pyriform cortex beta-waves: odor-specific sensitization following repeated olfactory stimulation. Brain Res 892: 301–308, 2001.[CrossRef][Web of Science][Medline]

Wiebe SP, Staubli UV. Dynamic filtering of recognition memory codes in the hippocampus. J Neurosci 19: 10562–10574, 1999.[Abstract/Free Full Text]

Wiltgen BJ, Brown RA, Talton LE, Silva AJ. New circuits for old memories: the role of the neocortex in consolidation. Neuron 44: 101–108, 2004.[CrossRef][Web of Science][Medline]

Wood ER, Dudchenko PA, Eichenbaum H. The global record of memory in hippocampal neuronal activity. Nature 397: 613–616, 1999.[CrossRef][Medline]

Yoon T, Otto T. Differential contributions of dorsal vs. ventral hippocampus to auditory trace fear conditioning. Neurobiol Learn Mem 87: 464–475, 2007.[CrossRef][Web of Science][Medline]

Zibrowski EM, Vanderwolf CH. Oscillatory fast wave activity in the rat pyriform cortex: relations to olfaction and behavior. Brain Res 766: 39–49, 1997.[CrossRef][Web of Science][Medline]

This article has been cited by other articles:

				D. A. Wilson Single-Unit Activity in Piriform Cortex during Slow-Wave State Is Shaped by Recent Odor Experience J. Neurosci., February 3, 2010; 30(5): 1760 - 1765. [Abstract] [Full Text] [PDF]

				J. Chapuis, S. Garcia, B. Messaoudi, M. Thevenet, G. Ferreira, R. Gervais, and N. Ravel The Way an Odor Is Experienced during Aversive Conditioning Determines the Extent of the Network Recruited during Retrieval: A Multisite Electrophysiological Study in Rats J. Neurosci., August 19, 2009; 29(33): 10287 - 10298. [Abstract] [Full Text] [PDF]

				A. Fontanini and D. B. Katz Behavioral States, Network States, and Sensory Response Variability J Neurophysiol, September 1, 2008; 100(3): 1160 - 1168. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 98/4/2196    most recent 00524.2007v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Martin, C. Articles by Kay, L. M. Search for Related Content PubMed PubMed Citation Articles by Martin, C. Articles by Kay, L. M.