| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
| |||||||
| |||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||
Department of Neurobiology, The Weizmann Institute of Science, Rehovot 76100, Israel
Submitted 1 March 2004; accepted in final form 13 June 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
| TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
). Attempts have been
made to identify and decode the neural mechanisms that subserve
familiarity of sensory inputs (Berman et al. 2000; Brown and Bashir 2002; Holscher et
al. 2003; Leveroni et
al. 2000; Li et al.
1993;
Tulving et al. 1996; Vankov et al. 1995). The study of
familiarity raises several experimental challenges. Familiarity
cannot be gauged on an absolute scale, and it depends on the sensory
history of the subject. Taste provides a convenient sensory
stimulus for the study of familiarity. Tastants are conveniently
controllable, and incidental exposure to a novel tastant is a salient
experience, which induces the formation of a long-term memory of that
tastant (Berman et al. 1998; Bures et al. 1998).
The rat gustatory cortex (GC) has been implicated in the acquisition,
consolidation, and retention of taste familiarity and taste
associations (Kiefer and Braun 1977; Rosenblum et al. 1993; Yamamoto et
al. 1994). Spiking
activity of GC neurons reflects both chemosensory and somatosensory
information (Hanamori et al. 1998; Yamamoto et al. 1989); these two components
can be partially separated along the temporal dimension of the
neural response (Katz et al. 2001). It has further been
reported that exposure to a novel tastant, as opposed to a familiar
one, activates multiple neurotransmitter and intracellular signal
transduction systems in the GC and that this activation is obligatory
for the encoding and retention of long-term taste memory (Berman
et al. 2000; Ferreira et al. 2002; Guiterrez et al. 1999; Rosenblum et
al. 1997). Activation of
the aforementioned systems outlasts taste presentation, develops over
a time scale of minutes, and persists for hours. It is not yet known
which neuronal events trigger these long-lasting processes and how
the encoded information is related to the familiarity of the tastant
processed during active stimulus sampling. It is plausible to assume
that the spiking activity of GC neurons, known to process
taste-related information within a fraction of a second (Halpern
1985),
subserves familiarity processing at the time of stimulation and
contributes to the immediate behavioral response to the taste.
Here, we examined how spiking responses of GC multi-units evolved
as rats became familiar with a novel taste stimulus. Tastants
were available for only 1 s during each trial to potentially
allow for better separation of somatosensory and chemosensory
inputs. Our results show that GC units signal taste familiarity
at a delayed temporal phase of the response. Our analysis suggests
that specific neuronal populations participate in the processing
of familiarity for specific tastants. Furthermore, the neural
signature of familiarity was correlated with familiarization
with a specific tastant rather than with any tastant. This signature
was not evident during the first session of exposure to the
novel tastant (lasting 
45
min), but rather only in a second session, 24 h later. Thus
persistent cortical representation of taste familiarity requires slow
post-acquisition processing to develop.
|
METHODS |
|---|
| TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Male Wistar rats (9–10 wk old, 250–300 g) were caged individually at 22 ± 2°C in a 12-h light/dark cycle, with food ad libitum. Ten rats participated in recording of neural activity; 24 rats participated in behavioral assessment of taste familiarity. All experiments were approved by the Animal Care and Use Committee of the Weizmann Institute of Science.
Behavioral assessment of taste familiarity
We determined behavioral familiarization to a tastant by using a
protocol of latent inhibition (Lubow 1989) of conditioned
taste aversion (CTA). This protocol quantifies familiarity in
terms of the decremental effect of pre-exposure on the ability
of a taste to enter into association in subsequent CTA training
to the same taste. Rats that were not subjected to recording
sessions (n = 6 in each group) were subjected once, or twice,
to a session of tastant drinking (sucrose, 5% wt/vol), under
the same experimental setup and conditions used during recordings
(see below). Two additional control groups were subjected to
the same protocols, but water was used instead of a tastant.
Twenty-four hours after the last session, all the rats were
trained in a CTA protocol as described in Bahar et al. (2003). Briefly,
they were presented with the tastant, and 40 min later, they were
injected intraperitoneally with the malaise-inducing agent, LiCl
(0.15 M, 2% body weight). CTA learning was quantified 48 h later in a
10-min multiple-choice test, in which the rats were allowed free
access to an array of six pipettes: three with 5 ml sucrose each and
three with 5 ml water each. Fluid consumption during the test was
measured, and an aversion index was calculated. Aversion index was
defined as {[ml of water/(ml of water + ml of sucrose)] x 100} consumed during the test. Hence,
an aversion index >50 implies preference of water over taste.
Recording settings
Training and recording sessions were conducted in a double-wall sound attenuating chamber (I.A.C.). The test cage (transparent Plexiglas; 42 x 25 x 30 cm) was fully covered with black cloth apart from one wall to allow a video camera, situated outside the cage, to monitor the rat's activity. A small amount of wood shavings from the home cage was placed in the training and recording cage to increase contextual familiarity. An optical lickometer (E24-01, Coulbourn Institute) and a multi-barrel pipette that allowed the delivery of multiple tastants were placed behind a 4.5 x 4.5 cm2 opening in the wall. Each barrel was connected to 0.2-mm-ID plastic tubing, connected to a 10-ml pipette via an electric pin valve (Angar Scientific). The opening in the cage wall was hidden behind a plastic gate driven by a small servoengine (HS-300, Hitech). A PC controlled and monitored the inputs from the optical lickometer and the openings of the valves and gate at a rate of 1 kHz. Another PC recorded neural activity. The two computers were synchronized via parallel ports.
Behavioral procedures
Rats were trained on the cued drinking task, described below, 1 wk
before and 1 wk after implantation of the electrodes. At the
beginning of training, rats were water deprived in their home cages
for 24 h. In the subsequent 5 days, they were transferred to the test
cage for 45–60 min daily. In the test cage, they were handled by the
experimenter (e.g., a light touch by the experimenter hand, 2 min) and trained to perform the
drinking task with water as the stimulus (150 trials daily). Each trial
started with an auditory cue that signaled the opening of the
gate (800 Hz, 100 ms); the rat was then given an opportunity to
lick the pipette. The optical lickometer monitored initial contact
with the pipette, as well as licking patterns. On initiation of
drinking, three aliquots of liquid were delivered (total volume, 50 ±
5 µl) over a period of 900 ms. At the end of this period, another
auditory cue (3,000 Hz, 100 ms) signaled the rat to remove its head
from the pipette. Immediately following the cue, the gate was partly
closed for 200 ms, preventing the rat from continuation of drinking
but allowing it to remove its head, and then it was closed
completely. The next drinking trial followed 10–12 s later
(randomly). It took the rat 2–3 training days to learn the task, at
which time it showed clear signs of anticipation following the first
auditory cue (e.g., moving in the direction of the gate) and
removed its head from the lickometer on the second auditory cue.
Following 5 days of training, the rat had free access to water for
3–4 days, after which it was implanted with chronic electrodes
(see Implantation of electrodes). Following recuperation from
surgery (7–14 days), the rat was water deprived for 24 h and
trained again for 4 days as described above. On the fifth day, it
performed 150 drinking trials while the head-stage was attached to
the electrode to familiarize it with the recording apparatus.
Implantation of electrodes
Rats were deeply anesthetized with sodium pentobarbital (85 mg/kg,
ip). Additional doses of pentobarbital were applied during surgery if
needed. Rats were restrained in a stereotaxic apparatus (David Kopf
Instruments, Tujunga, CA) with blunt ear bars. The skull was exposed,
and a small opening was made just above the right insular cortex (IC;
anteroposterior, +1.2 mm, lateral, +4.8–5.2 mm, relative to bregma;
Paxinos and Watson 1998). An
additional four openings were made at different locations on the
skull to allow insertion of screws that secured the dental cement and
served as ground. Multi-wire electrodes were made of eight-stainless
steel Teflon coated microwires (50 µm diam; impedance, 200–500 kOhms,
at 1 kHz), arranged in a bundle (150–250 µm diam; NB-labs, Dennison,
TX; Nicolelis et al. 1997). The electrode was
placed just above the IC, and the ground wire was attached to the
ground screws. The electrode was lowered slowly at a rate of
100–200 µm/1–5 min, until the IC was reached (4.7–5 mm
below dura). In two rats, an additional two wires (120 µm each) were
implanted in the right masseter muscle to record EMG activity during
the experiments. After final placement of the electrode, acrylic
dental cement was applied around the plastic connector to secure the
electrode to the skull. Rats were injected with 0.2 ml of Pen-Strap
to prevent infection. They were allowed 7–14 days of recuperation
with food and water available ad libitum.
Recording sessions
Recording experiments commenced immediately after the last training
session. Each experiment lasted 1 wk, with an inter-experiment
period of 1 wk. Each experiment was dedicated to one tastant.
NaCl (0.155 M) and sucrose (0.146 M) were used as tastants.
Water was double-distilled, either when used to dilute the tastants
or as a stand-alone stimulus. NaCl and sucrose were selected
because they do not evoke the motor activity typical to neophobic
reactions (Grill and Norgren 1978; Bahar and Dudai,
unpublished data). Six rats were subjected to two recording
experiments, one for each tastant stated above. Four additional rats
were subjected to only one recording experiment (due to
technical considerations). Each experiment included three tastant
sessions, designated NOVEL, FAMILIAR1, and FAMILIAR2, which
corresponded to the first, second, and third exposure to the same
tastant, respectively. These sessions were performed 24 h apart.
Each tastant session included four alternating blocks of water
(W) or tastant (T) application (designated W1, T1, W2, T2, and
given in that order). Each block consisted of 35 drinking
trials, so that overall, each session consisted of 70 trials to
water and 70 trials to tastant (Fig.
1A). Each session lasted 40–50 min. Two squirts of water
(200 µl each) flushed the drinking pipette at the end of each block.
Recordings from two additional sessions of water only were performed
before and after the aforementioned recording sessions, bringing the
total number of recording sessions in each experiment to seven. Each
water session was divided into four blocks (W1-4; see Fig.
1A) that were later analyzed as the blocks in tastant
sessions (see Data acquisition and analysis). Following each
session, rats were given 30 min to consume their daily water ration;
between experiments, they were given free access to food. Since rats
manage to consume only <50% of their daily water ration (7/15 ml) during each
recording session, we did not expect a major decline in the
level of thirst during recordings; this was also evident behaviorally
because rats showed high motivation to drink at the home cage
following each session.
|
During recordings, the electrodes were connected to a voltage-follower
head stage (NB-Labs). The analogue-recorded signals were amplified
and filtered (8,000–15,000x gain; 0.5-
to 4.5-kHz band-pass filter; MCP plus, Alpha-Omega, Nazareth,
Israel). The signals were digitized at a sampling rate of 30
kHz/channel (DAP 5200, Microstar Laboratories), displayed on a
computer screen, and stored using Alpha-Map acquisition program
(Alpha-Omega). To monitor facial motor activity, we recorded the EMG
of the facial masseter muscle, one of the main muscles responsible
for mouth movements during ingestion. EMG activity was sampled at a
rate of 0.5 kHz (2- to 200-Hz band-pass filter) and quantified
by calculating the number of times activity crossed 2.5 x SD of the inter-trial interval activity
(this threshold captured activity that deviated from the background
noise with a confidence of >98%). EMG values varied between rats,
and to combine them, we normalized the data points to the highest
value before averaging. We also monitored oro-facial activity using a
video monitor. Data were collected from recording experiments that
showed an increase in taste/water ratio during the LP in two rats.
The following typical oro-facial behaviors (Grill and Norgren 1978) were counted
following the end of each drinking trial: mouth movements lasting
<2 or >2 s; tongue protrusions lasting <2 or >2 s; chin
rubbing 
5 s.
Spike waveforms were sorted off-line using ASORT program (Alpha-Omega)
that employed principle component analysis (see Fig. 3 for
examples of sorted units). Standard methods of analysis were used
(Sosnik et al. 2001). Peristimulus time
histograms (PSTHs) were constructed at 1-ms resolution and smoothed
by a low-pass filtering cutoff frequency of 10 Hz. Most recorded
units showed a lingering response (above spontaneous firing rate) to
1 s of drinking stimulation, which lasted 7 s after drinking onset. We divided the
response into three temporal phases: early phase (EP), 0–1 s
after drinking onset; middle phase (MP), 1–2 s after drinking
onset; and late phase (LP), 2–7 s after drinking onset (see
PSTHs in Fig.
3). This division was based on the following rationale: EP was
the period in which the stimulus was presented. In addition, the
neural responses and EMG recordings of the facial muscle during this
phase showed strong rhythmic activity at 5–10 Hz (Fig. 4),
typical for licking activity (Katz et al. 2001); MP was the
poststimulus period in which 5- to 10-Hz oscillations were still
evident in the EMG recording and neural responses; in LP, oro-facial
activity declined markedly, but neural activity was still above
pretrial level (Fig. 4).
All further analyses were performed according to these response
phases.
|
|
To investigate responses across the whole population, an average spike count for each phase and for each trial was calculated based on unit-by-unit spike counts. Data from water blocks (W1, W2) or taste blocks (T1, T2) were combined, respectively. To analyze "water sessions," spike counts were calculated for each phase as above. In this analysis, blocks W2 and W4 during the "water sessions" were considered analogous to taste blocks during taste sessions (T1, T2), and therefore their data were combined. Blocks W1, W3 during the "water sessions" were considered analogous to water blocks in the taste sessions, and their data were also combined (see Fig. 1A).
Graphs that describe the distribution of taste/water ratios were constructed by dividing the average response to taste by that of water on a unit-by-unit base. The resulting distributions were usually not normal and therefore compared by Mann-Whitney U test. Exponential smoothing of the distributions was used for visualization (damping factor = 0.3).
PSTHs representing population responses (e.g., Fig.
8A) were constructed from the averaged activity of
multiple multi-units and smoothed by a convolution with a low-pass
zero-phase Butterworth filter with cutoff frequency at 2 Hz. This low
smoothing time constant was chosen to emphasize the effects at the
slow time scales, typical for the taste system (Halpern 1985). Data points
in the graphs that describe response dynamics along the sessions
were constructed by averaging six trials for each. After testing
several bin sizes, we selected this bin size as the best in
representing the data.
|
Histology
Following the last recording session, rats were deeply anesthetized
with sodium pentobarbital. A DC current was passed through each
wire (35 µA, 7 s) to induce a lesion. Next, rats were perfused
with saline and then with 5% formalin. The brains were removed and
fixed overnight with 5% formalin and 30% sucrose. Brain sections (40
µm) were labeled with Nissel-stain and examined to locate electrode
tips position (Haidarliu et al. 1999).
|
RESULTS |
|---|
| TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Figure
1B shows a representative lesion made by the tips of a
multi-wire bundle electrode (8 wires), located in the dysgranular
zone of the insular cortex (DI). Other recordings were also
made in this area. This area is known to contain cells that
respond to taste stimulation (Katz et al. 2001; Yamamoto et
al. 1989), and its functional
ablation impairs taste learning (Berman et al. 2000). Bundle wire tips were
usually spread across several cortical layers.
Behavior
To quantify behavioral familiarization with a taste stimulus under
our experimental protocol (see Fig.
1A), we used a protocol of latent inhibition (Lubow 1989) of CTA. In this
protocol, familiarization with a taste stimulus reduces its
associability in subsequent CTA training; thus the degree of
reduction in CTA performance can serve as a measure of the relative
familiarity by which the taste stimulus was perceived during training
(Rosenblum et al. 1997). Rats that were
pre-exposed to sucrose under the same conditions used in recordings
(but not subjected to recording themselves), showed a significant
decrease in CTA performance [aversion index, 71 ± 8 (SE) and 62 ± 4
for 1 and 2 pre-exposures, respectively) compared with rats that
were pre-exposed to water alone (94 ± 2; ANOVA,
F(2,15) = 12, P < 0.001; the difference
between the aversion indices of groups preexposed once or twice to
sucrose was not significant, P = 0.4, Tukey test; Fig. 2).
|
The neural data were obtained from 10 rats. Six rats participated in 2 experiments each, and four rats participated in only 1 experiment; hence data were used from 16 experiments altogether. Following off-line sorting, most signals were identified as belonging to multi-units, which will be referred to herewith as units (based on spike shapes and inter-spike interval analysis; only 4% were clear single units; see Fig. 3). No attempt was made therefore to trace units from 1 day to another, and analysis was at the population level throughout. For the NOVEL and FAMILIAR1 sessions, 169 units were recorded for each session. In 10 of 16 experiments (129 units), FAMILIAR2 sessions were also performed. Spontaneous firing rates were generally low, with an average of 0.92 spikes/s and a median of 0.74 spikes/s; rates ranged from 0.01 to 4.8 spikes/s. Mean spontaneous firing rate did not differ significantly between NOVEL, FAMILIAR1, and FAMILIAR2 sessions (ANOVA, F(2, 464) = 1.1, P > 0.2).
Analysis of taste responding units
We calculated the percentage in the total population of units whose evoked responses were significantly different between tastant and water (whether responses were higher or lower to a tastant compared with water). These units were dubbed TRUs. Basing our calculations on the response during the LP yielded the highest percentage of TRUs under all conditions. Percentages of TRUs in the LP were 27% (45/169 units), 36% (60/169 units), and 34% (44/129 units) for NOVEL, FAMILIAR1, and FAMAILIAR2 sessions, respectively (Fig. 5). However, Wilcoxon signed-ranks test showed no significant statistical differences in the percentages of TRUs between sessions during any of the phases (e.g., LP: NOVEL-FAMILIAR1: P = 0.27; NOVEL-FAMILIAR2, P = 0.95; FAMILIAR1-FAMILIAR2, P = 0.9).
|
We next calculated the spike count for each phase averaged across the total population of recorded units, regardless of their response to taste or water. Figure 6 shows that, while the average spike count was not significantly different between taste and water in the NOVEL session during any phase, there was a significant increase in average spike count for taste compared with water in the FAMILIAR1 and FAMILIAR2 sessions during the LP only [average response to water vs. tastant, 6.9 ± 0.3 vs. 7.1 ± 0.3, 7 ± 0.4 vs. 8.3 ± 0.4, 6.6 ± 0.2 vs. 7.7 ± 0.3 spikes/phase; paired t-test, P = 0.48, P < 0.001, and P < 0.001 in NOVEL (n = 169), FAMILIAR1 (n = 169), and FAMILIAR2 (n = 129), respectively]. As a control, we performed a similar spike count analysis on neural responses recorded during "water sessions" only (conducted before and after the NOVEL and FAMILIAR2 sessions, respectively; see Fig. 1A). This analysis showed that the average activity of the entire population was highly similar in the different blocks, within and across sessions during all phases, when rats drank water only (Fig. 6, insets). These data suggest that the increased response to a tastant during the familiar sessions was not a result of unstable recordings or the order of tastant and water blocks within a session.
|
|
We also tested whether, for each rat, the recorded population that showed a familiarity response to a specific tastant during the LP tended to respond specifically to this tastant during the EP and MP, regardless of familiarity. No significant correlation was found between the familiarity response to a specific tastant and the tendency of each population to respond specifically to this tastant during the EP or MP (sucrose, r2 = 0.06, P = 0.5; NaCl, r2 = 0.09, P = 0.4).
Analysis of response dynamics
To explore the dynamics of the response to the familiar tastant, we averaged PSTHs taken from experiments in which the recorded population showed a significant increased response to familiar tastant (94 units). Figure 8A shows that the difference between the averaged response to familiar taste versus water was relatively stable during the entire LP (in FAMILIAR1 and FAMILIAR2 sessions; data of the latter not shown). To explore the dynamics of the response during the LP throughout the entire session, we averaged spike counts taken from groups of six trials within sessions (grouping by 5 or 7 trials yielded similar results). Figure 8B shows that the increased spike count to familiar taste during the LP was consistent along the entire FAMILIAR1 session and that no earlier increase of average activity could be observed during the NOVEL session. We also analyzed the response dynamics along the sessions for each unit individually. Along the NOVEL sessions, most units displayed a tendency to decrease their response to water and tastant stimulations [water, slope = –0.35 ± 0.29 (SD) spikes/block of 6 trials; taste, slope = –0.27 ± 0.48 spikes/block of 6 trials). During the FAMILIAR1 sessions, units also tended to decrease their response but to a lesser degree (water, slope = –0.07 ± 0.22 spikes/block of 6 trials; taste, slope = –0.17 ± 0.42 spikes/block of 6 trials). The average slope did not differ significantly between the response to taste and water in each session (P > 0.1). We also tested whether the balance between units that increased or decreased their tastant/water response ratio ("excited" and "inhibited," respectively) remained constant during the NOVEL session. No change in the ratio of tastant/water response could be detected during the NOVEL session when units were divided into "excited" and "inhibited" by the tastant. The slopes of the tastant/water response ratios of the "excited" units were 0.08 ± 0.026 and those of the "inhibited" units were 0.05 ± 0.024, with no significant difference between them (P = 0.254).
Analysis of oro-facial activity
We monitored and analyzed several parameters related to oro-facial motor activity that could potentially affect the recorded neural responses. Figure 9A shows that rates of typical orofacial movements related to ingestion, monitored by a video camera during the experiments, were not significantly different between NOVEL and FAMILIAR1 sessions or between water and tastant blocks. We also found that the average number of licks per trial (during the EP) was similar among all sessions (7 ± 0.3 licks on average, data not shown). EMG recordings of the masseter muscle provided a more direct measure of oro-facial activity. Figure 9B shows that EMGs recorded from two rats did not change significantly between exposures to water or tastant during the LP in either the NOVEL or FAMILIAR1 sessions. In contrast, the average spike count, recorded from the same rats and during the same experiments, increased significantly in response to the familiar tastant compared with water in the FAMILIAR session (Fig. 9B). To further explore the relevance of the facial muscular activity to the neural response, we calculated the correlation between the EMG activity and the neural response on a trial-by-trial basis throughout the recording sessions and for each phase. The highest values of Pearson correlation coefficients were found during the EP (r2 = 0.5, P < 0.01) compared with the MP and LP (Fig. 9C; r2 = 0.18 and 0.13 in MP and LP, respectively; P > 0.2 for both).
|
|
DISCUSSION |
|---|
| TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Although we did not monitor in detail the emergence of the signature
of familiarity over time, the fact that it became apparent only
at 24 h but not at 45 min after the initial encounter with the
taste raises the possibility that the emergence of signature of
familiarity is a slow process, which might be related to the
activation of neurotransmitter receptors, modulation of gene
expression, and posttranslational modifications detected in the
insular cortex in the first hours after the consumption of an
unfamiliar, but not a familiar tastant (Berman et al. 1998, 2000; Gutierrez et al. 2003; Koh et al. 2003; Rosenblum et
al. 1997).
The rationale for our experimental protocol was based on two major
observations. First, a single session of drinking a novel tastant
elicits long-term taste memory (Berman et al. 2000; Bures et al.
1998). Second,
consolidation of taste memory depends on neural mechanisms active
over minutes to hours (e.g., Berman et al. 1998; Rosenblum et al. 1993). Hence, rats were
presented with the same tastant when it was novel (NOVEL session),
when it became familiar 1 day later (FAMILIAR1 session), and at
48 h (FAMILIAR2 session). This allowed distinct dissociation in
time of processes that subserve memory acquisition and consolidation
(NOVEL session) and memory retrieval (FAMILIAR sessions). The
presentation of alternating blocks of repeated water and tastant
(Fig.
1A) allowed us to ensure sufficient and continuous exposure
to the stimuli, while at the same time, reducing the potential
for sensory adaptation.
It is noteworthy that previous data have indicated that a few
licks of a new tastant are insufficient to render that tastant
familiar, as judged by the ability to form taste associations
in CTA as well as to hinder association in subsequent CTA in a
latent inhibition protocol; rather, it takes a substantial amount of
consumed tastant, comparable to that consumed in the NOVEL session,
to effectively induce long-term taste memory (Rosenblum et al. 1997; Stehberg and Dudai,
unpublished data). Furthermore, both behavioral and molecular data
show that it takes many minutes to several hours to establish taste
familiarity (Berman et al. 1998, 2000; Rosenblum et al. 1997). We therefore
assume that, when acquisition and consolidation of lasting taste
familiarity are considered, one could treat the trials in the
NOVEL session as repetitive exposures to a yet unfamiliar taste.
Responses in the GC are driven by converging somatosensory and
chemosensory inputs arriving from the oral cavity and detected
within seconds (Hanamori et al. 1998; Kosar et al. 1986; Yamamoto et
al. 1989). In some
studies, durations of each stimulus exposure were relatively long
(often 5–20 s). This complicated the separation between somatosensory
and chemosensory inputs. To allow for at least partial separation
among these input variables, we restricted exposure to the stimuli to
1 s in each trial. In addition, we analyzed the response in three
separated phases determined by stimulus duration and oro-facial motor
activity (Figs. 3 and 4). In all our
analyses, response to water stimulation was used as a reference to
the response to the tastant, since we considered water to be a highly
familiar stimulus. This was also evident in the stable responses to
water stimulations during water sessions (Fig. 6,
insets).
We first employed an analysis that considered only the units that
showed significantly different response to the tastant (TRUs). This
resembles a commonly employed analysis, which showed that relatively
few cells responded to a specific tastant (10–15%; Hanamori et al. 1998; Yamamoto et al. 1989; Yasoshima and
Yamamoto 1998). Our results from
analyzing the responses in the EP (Fig. 5)
are consistent with these results. However, consistent with
Katz et al. (2001), we show that the
percentage of TRUs is larger when considering additional responses
phases. In particular, we found 27% of TRUs when analyzing the LP of
the NOVELL session.
We found that familiarity was correlated with increased activity to a tastant during the LP when the entire recorded population in the GC was considered (Fig. 6). Examination of this increased LP activity on a session by session basis (Fig. 7), suggested that specific populations participated in the processing of familiarity for specific tastants, and that the response to a familiar tastant during the LP was not a general response to familiar stimuli, but rather, correlated with the familiarization to a specific experienced tastant. Thus the percentage of activity change measured for the entire GC population should be taken as a lower bound for the "signal-to-noise" estimated for a presumed taste-specific, and perhaps familiarity-specific, neuronal group. In addition, the analysis that found no significant correlation between the familiarity response to a specific tastant and tastant specific response during the EP and MP suggests that the populations that encode familiarity of a tastant do not necessarily encode the familiarity-independent attributes of this tastant.
The increased activity during the LP in the entire population (Fig. 6) and
the variability in TRUs between the phases (Fig. 5)
are in accordance with reports that single neurons may process
multiple types of information in the time course of their response
(Lipton et al. 1999; McClurkin and Optican
1996; Sugase et
al. 1999). In particular, it has
been shown that single units in the GC can dynamically process
chemosensory and somatosensory information along their time-varying
responses (Katz et al. 2001, 2002a). Further support for
this hypothesis stems from the analysis that correlated the time
course of the EMG activity with neural response: a significant
correlation was observed in EP only (Fig.
9C), suggesting that somatosensory inputs are processed
mainly during the EP. Similar multi-modal processing phases were
observed previously, although on a faster time scale (Katz et al.
2001, 2002a). The difference in time
scale might result from the difference in the experimental protocol
of the two studies. A proposed scheme of the dynamic processing
of different variables in the GC is depicted in Fig. 10.
|
) had similar
probabilities of occurrence and duration in the NOVEL and FAMILIAR1
sessions (Fig.
9A). Second, EMG recordings of the masseter muscle, a
major facial muscle that controls jaw movements (Yamamoto et al.
1982), did
not unveil differences in muscle activity between the different
sessions. In contrast, neural activity recorded from rats that
participated in the EMG recordings showed a significant increase
in response to the familiar tastant (Fig.
9B). Third, if the increased response to a familiar
tastant was indeed the result of increased motor reaction to a
familiar stimulus, we would have expected an increase in neural
response to all familiar tastants. However, rats tested on two
different tastants showed increased response to one only (Fig. 7). This
was detected in experiments using either one of the tastants, hence
is unlikely to reflect familiarization with a unique taste quality.
It is noteworthy that the taste cortex occupies 4.5 mm2 (Bures et
al. 1998), whereas our electrode
bundle captured neural responses from a volume 500-fold smaller. Thus the probability of
detecting familiarity signal to both tastants by recording from one
bundle was inherently low. This low probability is also in line
with the assumption that, in the taste cortex, stimulus
properties are encoded nonhomogeneously over the cortical surface
(Yamamoto et al. 1989).
What might be the physiological basis of the persistent activity
observed in the GC during the LP? This activity is not the sum
response of early and late response single units, since responses
of single units in our recordings also contained LP (Fig.
3C; Katz et al. 2001). It is also unlikely
that strong somatosensory inputs affect activity during the LP, as
evident from the low oro-facial activity measured during this phase
(Fig. 4);
inputs that may rise from swallowing activity are also unlikely,
because it was shown that the rat swallows small amounts of liquid
fast (Travers and Norgren 1986). Although we cannot
rule out the possibility that the LP may be affected to some degree
by somatosensory inputs, we prefer the interpretation that GC
persistent responses arise from interactions with other stations in
the taste system (e.g., amygdala, Escobar and Bermudez-Rattoni 2000; van der Kooy
et al. 1984) or reverberations of
intra-cortical activity (Katz et al. 2002b). Our data show that the
increased response to a familiar tastant was stimulus specific, thus
the increased response during the LP could reflect retrieval of
stored representation of specific familiar tastants, suggested to
reside in the GC (Berman et al. 2000).
All in all, an attractive hypothesis is that networks in the GC
function as associative memory networks (Hopfield 1982), which learn
new tastants according to Hebbian-like rules under neuromodulatory
control (Ahissar et al. 1992, 1996, 1998; Barkai and
Hasselmo 1997; Ego-Stengel et al.
2001; Shulz et al.
2000). On
the system's algorithmic level, is formation of taste memory
triggered by taste novelty or inhibited by taste familiarity
(Berman et al. 2000)? Since the spontaneous
and evoked activity of GC neurons is relatively low, it is tempting
to assume, on the basis of our data, that the transformation of
representation of novelty to familiarity is signaled by increased,
rather than decreased, activity; this suggests that the cortex, as
default, considers a tastant novel until proven otherwise. Hence,
the formation of taste memory in the GC might be expected to be
inhibited by taste familiarity (Fig. 11).
This seems useful from a phylogenetic point of view, as instantaneous
consumption of a novel foodstuff may prove lethal.
|
|
GRANTS |
|---|
| TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
| TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: E. Ahissar, Dept. of Neurobiology, The Weizmann Institute of Science, Rehovot 76100, Israel (E-mail: ehud.ahissar{at}weizmann.ac.il ).
|
REFERENCES |
|---|
| TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Ahissar E, Haidarliu S, and Shulz D. Possible involvement of neuromodulatory systems in cortical Hebbian-like plasticity. J Physiol Paris 90: 353–360, 1996.[CrossRef][ISI][Medline]
Ahissar E, Vaadia E, Ahissar M, Bergman H, Arieli A, and Abeles M. Dependence of cortical plasticity on correlated activity of single neurons and on behavioral context. Science 257: 1412–1415, 1992.[ISI][Medline]
Bahar A, Samuel A, Hazvi S, and Dudai Y. The amygdalar circuit that acquires taste aversion memory differs from the circuit that extinguishes it. Eur J Neurosci 17: 1527–1530, 2003.[CrossRef][ISI][Medline]
Barkai E and Hasselmo MH. Acetylcholine and associative memory in the piriform cortex. Mol Neurobiol 15: 17–29, 1997.[ISI][Medline]
Berman DE, Hazvi S, Neduva V, and Dudai Y.
The role of identified neurotransmitter systems in the response of insular
cortex to unfamiliar taste: activation of ERK1-2 and formation of a memory
trace. J Neurosci 20: 7017–7023,
2000.
Berman DE, Hazvi S, Rosenblum K, Seger R, and
Dudai Y. Specific and differential activation of mitogen-activated protein
kinase cascades by unfamiliar taste in the insular cortex of the behaving rat.
J Neurosci 18: 10037–10044, 1998.
Bronson G. The fear of novelty. Psychol Bull 5: 350–358, 1967.
Brown MW and Bashir ZI. Evidence concerning how neurons of the perirhinal cortex may effect familiarity discrimination. Philos Trans R Soc Lond B 357: 1083–1095, 2002.[CrossRef][ISI][Medline]
Bures J, Bermudez-Rattoni F, and Yamamoto T. Conditioned Taste Aversion. Memory of a Special Kind. Oxford: Oxford University Press, 1998.
Ego-Stengel V, Shulz DE, Haidarliu S, Sosnik
R, and Ahissar E. Acetylcholine-dependent induction and expression of
functional plasticity in the barrel cortex of the adult rat. J
Neurophysiol 86: 422–437, 2001.
Escobar ML and Bermudez-Rattoni F. Long-term potentiation in the insular cortex enhances conditioned taste aversion retention. Brain Res 852: 208–212, 2000.[CrossRef][ISI][Medline]
Ferreira G, Gutierrez R, De La Cruz V, and Bermudez-Rattoni F. Differential involvement of cortical muscarinic and NMDA receptors in short- and long-term taste aversion memory. Eur J Neurosci 16: 1139–1145, 2002.[CrossRef][ISI][Medline]
Grill HJ and Norgren R. The taste reactivity test. I. Mimetic responses to gustatory stimuli in neurologically normal rats. Brain Res 143: 263–279, 1978.[CrossRef][ISI][Medline]
Gutierrez H, Gutierrez R, Ramirez-Trejo L,
Silva-Gandarias R, Ormsby CE, Miranda MI, and Bermudez-Rattoni F. Redundant
basal forebrain modulation in taste aversion memory formation. J Neurosci
19: 7661–7669, 1999.
Gutierrez R, Tellez LH, and Bermudez-Rattoni F. Blockade of cortical muscarinic but not NMDA receptors prevents a novel taste from becoming familiar. Eur J Neurosci 17: 1556–1562, 2003.[ISI][Medline]
Haidarliu S, Sosnik R, and Ahissar E. Simultaneous multi-site recordings and iontophoretic drug and dye applications along the trigeminal system of anesthetized rats. J Neurosci Methods 94: 27–40, 1999.[CrossRef][ISI][Medline]
Halpern BP. Time as a factor in gustation: temporal patterns of taste stimulation and response. In: Taste, Olfaction, and the Central Nervous System, edited by Pfaff DW. New York: Rockefeller University Press, 1985, p. 181–209.
Hanamori T, Kunitake T, Kato K, and Kannan
H. Responses of neurons in the insular cortex to gustatory, visceral, and
nociceptive stimuli in rats. J Neurophysiol 79: 2535–2545, 1998.
Holscher C, Rolls ET, and Xiang J. Perirhinal cortex neuronal activity related to long-term familiarity memory in the macaque. Eur J Neurosci 18: 2037–2046, 2003.[CrossRef][ISI][Medline]
Hopfield JJ. Neural networks and physical systems with emergent selective computational abilities. Proc Natl Acad Sci USA 79: 2554–2558, 1982.[Abstract]
Katz DB, Nicolelis MAL, and Simon SA. Gustatory processing is dynamic and distributed. Curr Opin Neurobiol 12: 448–454, 2002a.[CrossRef][ISI][Medline]
Katz DB, Simon SA, and Nicolelis MAL.
Dynamic and multimodal responses of gustatory cortical neurons in awake rats.
J Neurosci 21: 4478–4489, 2001.
Katz DB, Simon SA, and Nicolelis MAL.
Taste specific neuronal ensembles in the gustatory cortex of awake rats. J
Neurosci 22: 1850–1857, 2002b.
Kiefer SW and Braun JJ. Absence of differential associative responses to novel and familiar taste stimuli in rats lacking gustatory neocortex. J Comp Physiol Psychol 91: 498–507, 1977.[ISI][Medline]
Koh MT, Wilkins EE, and Bernstein IL. Novel tastes elevate c-fos expression in the central amygdala and insular cortex: implications for taste aversion learning. Behav Neurosci 117: 1416–1422, 2003.[CrossRef][ISI][Medline]
Kosar E, Grill HJ, and Norgren R. Gustatory cortex in the rat. I. Physiological properties and cytoarchitecture. Brain Res 379: 329–341, 1986.[CrossRef][ISI][Medline]
Leveroni CL, Seidenberg M, Mayer AR, Mead LA,
Binder JR, and Rao SM. Neural systems underlying the recognition of familiar
and newly learned faces. J Neurosci 20: 878–886, 2000.
Li L, Miller EK, and Desimone R. The representation of stimulus familiarity in anterior inferior temporal cortex. J Neurosci 69: 1918–1929, 1993.
Lipton PA, Alvarez P, and Eichenbaum H. Crossmodal associative memory representations in rodent orbitofrontal cortex. Neuron 22: 349–359, 1999.[CrossRef][ISI][Medline]
Lubow RE. Latent Inhibition and Conditioned Attention Theory. Cambridge: Cambridge University Press, 1989.
McClurkin JW and Optican LM. Primate
striate and prestriate cortical neurons during discrimination. I. Simultaneous
temporal encoding of information about color and pattern. J Neurophysiol
75: 481–495, 1996.
Nicolelis MAL, Ghazanfar BMF, Faggin BM, Votoaw S, and Oliveira LMO. Reconstructing the engram: simultaneous, multisite, many single neuron recordings. Neuron 18: 529–537, 1997.[ISI][Medline]
Paxinos G and Watson C. The Rat Brain in Stereotaxic Coordinates (4th ed.). New York: Academic Press, 1998.
Rosenblum K, Berman DE, Hazvi S, Lamprecht R,
and Dudai Y. NMDA receptor and the tyrosine phosphorylation of its 2B
subunit in taste learning in the rat insular cortex. J Neurosci 17:
5129–5135, 1997.
Rosenblum K, Meiri N, and Dudai Y. Taste memory: the role of protein synthesis in gustatory cortex. Behav Neural Biol 59: 49–56, 1993.[ISI][Medline]
Shulz DE, Sosnik R, Ego V, Haidarliu S, and Ahissar E. A neuronal analogue of state-dependent learning. Nature 403: 549–553, 2000.[CrossRef][ISI][Medline]
Sosnik R, Haidarliu S, and Ahissar E.
Temporal frequency of whisker movement. I. Representations in brain stem and
thalamus. J Neurophysiol 86: 339–353, 2001.
Sugase Y, Yamane S, Ueno S, and Kawano K. Global and fine information coded by single neurons in the temporal visual cortex. Nature 400: 869–873, 1999.[CrossRef][ISI][Medline]
Travers JB and Norgren R. Electromyographic analysis of the ingestion and rejection of sapid stimuli in the rat. Behav Neurosci 100: 544–555, 1986.[CrossRef][ISI][Medline]
Tulving E, Markowitsch HJ, Craik FIM, Habib R, and Houle S. Novelty and familiarity activation PET studies of memory encoding and retrieval. Cereb Cortex 6: 71–79, 1996.[Abstract]
van der Kooy D, Koda LY, McGinty JF, Gerfen CR, and Bloom FE. The organization of projections from the cortex, amygdala, and hypothalamus to the nucleus of the solitary tract in rat. J Comp Neurol 224: 1–24, 1984.[ISI][Medline]
Vankov A, Minvielle AH, and Sara SJ. Response to novelty and its rapid habituation in locus coeruleus neurons of the freely exploring rat. Eur J Neurosci 7: 1180–1187, 1995.[ISI][Medline]
Yamamoto T, Matsuo R, Fujiwara T, and Kawamura Y. EMG activities of masticatory muscles during licking in rats. Physiol Behav 29: 905–913, 1982.[CrossRef][ISI][Medline]
Yamamoto T, Matsuo R, Kiyomitso Y, and
Kitamura R. Taste responses of cortical neurons in freely ingesting rats.
J Neurophysiol 61: 1244–1258, 1989.
Yamamoto T, Shimura T, Sako N, Yasoshima Y, and Sakai N. Neural substrates for conditioned taste aversion in the rat. Behav Brain Res 65: 123–137, 1994.[CrossRef][ISI][Medline]
Yasoshima Y and Yamamoto T. Short-term and long-term excitability changes of the insular cortical neurons after the acquisition of taste aversion learning in behaving rat. Neurosci 84: 1–5, 1998.[CrossRef][ISI][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
