|
nif:isString
|
-
Male Sprague-Dawley rats (n = 10) were purchased from the Canadian Breeding Laboratories (St-Constant, QC, Canada). All the rats were housed in individual plastic home cages lined with wood shavings and maintained on a 12:12 dark-light cycle (with the lights on from 06:00 to 18:00), with an ambient temperature of 22 ± 1°C, and free access to tap water and the standard laboratory rat diet (Rat/Mouse/Hamster chow, 1000 Formula, 12.9 kJ/g; Agway Prolab, Syracuse, NY), unless otherwise specified. All rats were cared for and handled according to the Canadian Guide for the Care and Use of Laboratory Animals, and the present protocol was approved by our institutional animal care committee (Comité de protection des animaux de l'Université Laval).
Rats were anesthetized using 4% isoflurane mixed with 2% oxygen; the dose was gradually reduced to 2% isoflurane– 1% oxygen during the surgery. Under anesthesia, the head was shaved, eye lubricant was applied, the analgesic buprenorphine (0.02 mg/kg) was administered subcutaneously under the abdominal skin, and a mixture of lidocaine (7 mg/kg) and bupivacaine (3.5 mg/kg) was injected subcutaneously over the skull incision site and area in contact with the stereotaxic ear bars. A chronically implanted unilateral 25-gauge guide cannula was aimed at the dorsal border of the cAHA (1.32 mm caudal to the bregma, 0.6 mm lateral to the midline, and 8.6 mm ventral to the scalp) [30]. Bundles of 16 microwire electrodes (platinum-iridium, formvar coated, 25-μm diameter, with an impedance of 250–500 kΩ measured at 1 kHz; California Fine Wire, Grover Beach, CA), along with one 50-μm diameter reference electrode made of the same material and insulation (California Fine Wire, Grover Beach, CA) were lowered into the guide cannula and protruded 1 mm below the distal end of the guide cannula (Fig 1). A stainless steel wire, as the conventional ground, was shouldered on one of the supporting screws on the skull. Rats were hydrated with saline and subcutaneously administered with the anti-inflammatory analgesic meloxicam (1 mg/kg) after completion of the surgery and before they were returned to the home cages for 7 days of recovery.
Figure data removed from full text. Figure identifier and caption: 10.1371/journal.pone.0156563.g001 A, Thionine-stained coronal brain section at the level of the central nucleus of the anterior hypothalamic area (cAHA; 1.32 mm caudal to the bregma) depicting electric micro-lesions (indicated by the white arrow) at the tip of the recording electrode that was lowered 1 mm ventral from the distal edge of the guide cannula. B, Panel B is an enlargement of a square shown in panel A. SO, supraoptic hypothalamic nucleus. Scale bar, 200 μm.
Recording of sucrose-licking events and neuronal activity in the cAHA: After 7 days of post-operative care and 2 days of habituation to an operant chamber equipped with two photo-beam lickometers (Med Asscociates Inc, VT, USA) supplied only with water, the rats were given 1 h of daily access to 10% sucrose, in addition to water, in the operant chamber in the following 2 days, in order to counter neophobia to the taste of sucrose. After the following 2 days on ad libitum access to chow in their home cages, rats were given 1 h of access to 10% sucrose and water in the operant chamber, during which time simultaneous acquisition of licking activity and extra-cellular multi-unit recordings in the cAHA was performed using a custom-made interface between the operant chamber and TDT multichannel acquisition system (Tucker Davis Technologies, FL, USA). This recording session provided data on sucrose licking and multi-unit activity in control, non-stressful conditions. After recordings in the non-stressful conditions, the rats were given two days of ad libitum access to chow in their home cages. Thereafter, the rats were subjected to mild foot shock stress (0.6 mA, 3 s duration, 4 times with an inter-shock interval of 15 s). Immediately after the stress session, the rats were placed in the operant cage, where multi-unit and licking activity were recorded during the 1-h access to 10% sucrose and water. Bottles of sucrose and water were weighed before and after the 1-h access in the operant chamber to evaluate intake. For each rat one control recording session and one recording session after exposure to stress were performed. On completion of the experiment, the rats were deeply anesthetized (60 mg/kg ketamine and 7.5 mg/kg xylazine), and the locations of the recording electrodes were marked by passing an anodal direct current of 90 μA for 15 s through selected electrode pairs in each bundle. The rats were then euthanized under anesthesia (60 mg/kg ketamine and 7.5 mg/kg xylazine) by intracardial perfusion with saline followed by phosphate buffered 4% paraformaldehyde. The brains were removed and stored at 4°C in a phosphate buffered 4% paraformaldehyde solution before they were cut into 40-μm coronal brain sections and stained with thionin. Rats with correct electrode placement to the cAHA (n = 8) were used for statistical analysis.
Lick timestamps were used to perform lick microstructure analysis. The total number of licks was determined in each 1-h recording session. Because water intake was very low (from 0 ml to 1 ml) during the 1-h assess to sucrose, the licking microstructure was analyzed only for sucrose. Sucrose-licking clusters were defined as high-frequency (6–9 Hz) licks occurring in a run of three or more licks interrupted by pauses of inter-cluster intervals (ICIs) of 3 s or longer [31]. The total number of licks, cluster number, cluster duration (s) and number of licks per cluster were calculated for each 1-h session using a custom written MATLAB script (R2010a, The MathWorks™). The total meal duration (s) was calculated as the sum of all cluster durations during the 1-h access to sucrose. Lick frequency within a cluster was estimated by dividing the number of licks per cluster by cluster duration. The mean inter-lick interval (ILI, ms) was calculated by creating an inter-lick interval histogram of the total lick events during the 1-h recording session with a bin size of 1 s.
Multi-unit activity was recorded using a 25-kHz sampling rate and signal filtering between 300 and 8000 Hz. The units with a signal-to-noise ratio of 3:1 were discriminated using a digital window discriminator. The principal component analysis spike sorter (Tucker Davis Technologies, FL, USA) was used for online sorting of captured spikes and their storage for offline analysis. Open sorter (Tucker Davis Technologies, FL, USA) was used to further refine the recorded multi-unit activity into individual, well-defined single-unit clusters using k-mean clustering. The mean waveform shape of each unit was identified using NeuroExplorer (Nex Technologies, Dallas, TX, USA). To characterize the neurons according to sucrose licking activity, 6-s (±3 s, 100-ms bin duration) peri-event rasters and peri-event histograms (PEHs) were created by aligning the unit firing to the licking cluster start (CS) or cluster end (CE) as a reference event (Figs 2 and 3). The resulting PEHs were smoothed with a Gaussian filter (bin width = 3) to reduce fluctuation in the firing rate due to licking. The PEH bin values are given in spikes per second and represent the mean frequency across all licking clusters for each neuron.
Figure data removed from full text. Figure identifier and caption: 10.1371/journal.pone.0156563.g002 A, Example of simultaneous recording of sucrose-licking events and the neuronal firing of a neuron with an increased firing rate at the cluster start (CS), termed CS-excited neuron. Clustering of the sucrose-licking events was based on inter-cluster intervals (ICIs) ≥ 3 s. B, Peri-event histograms (PEHs) and rasters of a CS-excited neuron in non-stressful conditions. C, PEHs and rasters of a CS-excited neuron after exposure to stress. Note that the firing rate of the CS-excited neurons decreased at the end of the sucrose-licking clusters (CE). In the rasters, the red triangles indicate sucrose licks and the black vertical dashes indicate unit discharges. The bin duration in the PEHs was 100 ms. Bin frequency values (impulses per second) from –2 s to –1 s were used to calculate the mean baseline frequency. Bin frequency values from 0 s to +1 s were used to characterize response to the CS.
Figure data removed from full text. Figure identifier and caption: 10.1371/journal.pone.0156563.g003 A, Example of simultaneous recording of sucrose-licking events and the neuronal firing rate of a neuron with decreased firing rate at the cluster start (CS), termed CS-inhibited neuron. Clustering of the sucrose-licking events was based on inter-cluster intervals (ICIs) ≥ 3 s. B, Peri-event histograms (PEHs) and rasters of a CS-inhibited neuron in non-stressful conditions. C, PEHs and rasters of a CS-inhibited neuron after exposure to stress. Note that the firing rate of the CS-inhibited neurons increased at the end of the sucrose-licking cluster (CE). In the rasters, the red triangles indicate sucrose licks and the black vertical dashes indicate unit discharges. The bin duration in the PEHs was 100 ms. Bin frequency values (impulses per second) from –2 s to –1 s were used to calculate the mean baseline frequency. Bin frequency values from 0 s to +1 s were used to characterize response to the CS.
The mean baseline firing rate (shown in Figs 2B and 3B) of each neuron was determined in the –2 s to –1 s time interval before the first lick of the cluster. The mean firing rate during the response to CS was calculated from 0 s to 1 s (Figs 2B and 3B). The baseline and response mean firing rate of all the recorded neurons was calculated for each experimental condition. Similar to the CS analysis, we also calculated statistically significant changes by using CE as a reference event. However, the majority of neurons showed similar results for the CS and CE analyses. To simplify the categories in terms of neuronal response, the final analyses were performed using only CS as the reference event, and we reported data only from this analysis. The neurons were classified as CS-excited (excitation at CS, Fig 2) if 1 s after the CS the frequency counts (10 bin counts from 0 s to 1 s, with a 100-ms bin duration) were significantly higher (Wilcoxon signed-rank test, p< 0.05) than the frequency counts at the baseline (10 bin counts from –2 s to –1 s, with a 100-ms bin duration). The neurons were classified as CS-inhibited (inhibition at CS, Fig 3) if 1 s after the CS the frequency counts were significantly lower than the frequency counts at the baseline. If no significant difference was found between the frequencies after CS (0 s to +1 s) and at the baseline (–2 s to –1 s), the neuron was classified as CS-nonresponsive at the start of sucrose licking. A non-parametric test (Wilcoxon signed-rank test) was used because the normality test (Kolmogorov–Smirnov, p< 0.05) did not show normal distribution of the frequency counts. To estimate the general 1-h dynamics of the firing rate of CS-classified neurons, rate histograms with a bin value of 1 s were created for each neuron for the entire 1 h of recording, with the exception of the first 2 min. The first 2 min were excluded from this analysis because during the first 2 min, system calibration and establishment of the units’ thresholds across the recording channels at the beginning of each recording sessions were performed. The mean firing rate across the 1-h recording was calculated in non-stressful and stressful conditions for CS-classified groups of neurons. The correlation between the firing rate of CS-inhibited, CS-excited and CS-nonresponsive neurons in non-stressful and stressful conditions was assessed using the non-parametric Spearman correlation test.
A fraction of the AHA neurons showed burst activity in the in vitro recordings and in the in vivo recordings under anesthesia [32,33]. To the best of our knowledge, this study is the first to present the burst analysis of the AHA neurons in freely moving rats. To examine the burst activity of the cAHA neurons, we used the Poisson surprise method [34]. In the Poisson surprise method as implemented in NeuroExplorer (Nex Technologies, AL, USA), a burst is detected as a group of three or more spikes with an inter-spike interval (ISI) less than half of the average ISI in the whole spike train (ISI in the burst < the average ISI/2). For each selected burst, the Poisson surprise value, that is, the negative logarithm of the probability of the occurrence of a burst in a random (Poisson) spike train [34], was calculated. The following algorithm attempted to maximize the surprise value for the burst by adding the next intervals in the spike train and removing intervals from the beginning of the burst. We used a minimum burst surprise value of 10, with which a high level of statistical significance (P < 0.00005) can be ensured for burst detection in a spike train. The Poisson surprise method is not sensitive to fluctuations in the average firing rate and inter-burst ISI [34,35]. This method was successfully used for detection of burst activity in cortical and subcortical neurons [34–39]. For each unit, several properties of bursts were determined including the burst number in the whole spike train, percentage of spikes in the bursts compared to the spikes number in the whole train, burst duration, mean spike number in the bursts and ISI of the bursts. In addition, we investigated whether the burst start (BS) or burst end (BE) time of the cAHA neurons influenced the sucrose licking activity. We created 1-s PEHs (± 0.5 s, 100-ms bin duration) by aligning the licking events to BS or BE as a reference event. The Wilcoxon signed-rank test was used to find significant (p < 0.05) differences between the lick frequency before (5 bin counts from –0.5 s to 0 s, with a 100-ms bin duration) and after (5 bin counts from –0 s to 0.5 s, with a 100-ms bin duration) the reference event. To assess whether BS occurred simultaneously with CS, we created 500-ms (± 250 ms with a bin size of 1 ms) cross-correlograms between the BS and CS events. The higher level of the confidence interval was set to 95%.
All results are presented as mean ± SEM. The paired Student’s t-test (p < 0.05) was used to compare sucrose intake and variables of the sucrose licking microstructure in non-stressful and stressful conditions. To assess whether the distribution of neuronal firing rates, responsiveness to sucrose and burst parameter values were normal, the Kolmogorov-Smirnoff test (p < 0.05) was used. The non-parametric Wilcoxon signed-rank test was used to compare the baseline and response firing rate of neurons classified according to sucrose licking clusters, and to calculate the differences in lick frequency at the firing burst start and end. The firing rate around the CS (baseline and response) of the CS-inhibited, CS-exited, and CS-nonresponsive neurons was analyzed using 2-way repeated-measures ANOVA to detect the main and interactive effects of stress and response to CS. Differences between individual groups were assessed using the Bonferroni post-hoc test. The 1-h mean firing rate of the CS-classified neurons was compared using the Kruskal-Wallis test following by Dunn’s multiple comparisons test. The non-parametric Spearman correlation test was used to determine the correlation between pairs of CS-classified groups. A Spearman’s rank correlation coefficient (r) value of >0.5 was considered to indicate a significant correlation between the pairs. The bursts in non-stressful and stressful conditions were assessed using the non-parametric Mann-Whitney test. The Pearson chi-square test was used to compare the number of neurons classified according to cluster-related activity. Results were considered significant at P values <0.05. The statistical tests were performed using GraphPad (GraphPad Software Inc., La Jolla CA, USA), MATLAB (R2010a, The MathWorks™, Natick, MA) and NeuroExplorer (Nex Technologies, AL, USA).
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}