Short- and long-term habituation of auditory event-related potentials in the rat

Animals

Male Wistar rats (National Laboratory Animal Center, University of Eastern Finland, Finland, n=12, weight 412 ± 9 g) were reared in groups of 2–4 until 5 months of age and individually thereafter in a controlled environment (temperature +21°C, lights on from 7:00 h to 19:00 h, water and food available ad libitum) Animals were housed in stainless steel metal cages, floor 31 cm x 45 cm, height 18 cm as according to the guidelines of the Council of Europe ETS123. At the age of 5–6 months, the rats were chronically implanted with two recording electrodes made of 50 μm insulated stainless steel wire (California Fine Wire Company Co, Grover Beach, CA, USA) in the hippocampus at the following stereotactic coordinates: AP (from Bregma) - 3.8, L (from Bregma) +3.1, V (from brain surface) - 3.1 with a vertical separation of the tips of 0.6 mm. In addition, two cortical screw electrodes (Wurth Electronics, Finland) were fixed on the (left and right) parietal bones (L ± 2.0 mm and A -7.5 mm from Bregma). A frontal screw served as the ground and a common reference electrode. The hippocampal electrode closest to the pyramidal cell layer and the right parietal cortical electrode were selected for the final analysis of evoked potentials. The rats were anesthetized with a mixture of pentobarbital and chloral hydrate (40 mg/kg i.p. each), and, for post-operative analgesia, they received 5 mg/kg of carprofen (Rimadyl^®, Vericore, Dundee, UK) intraperitoneally. The rats were housed in individual cages after the surgery. Recordings started after at least 2 weeks of recovery period. The rats were involved in an EEG study for three weeks before the current study on evoked potentials. All animal procedures were carried out in accordance with the guidelines of the European Community Council Directives 86/609/EEC and approved by the Animal Experiment Board of Finland.

Data acquisition

In total 10 rats were recorded for the study but due to poor signal in some channels, the number of records in the analysis varies from 6 to 9. During the recordings the rat was able to freely move in a brown paste-board cylinder (70 cm diameter, 50 cm height) that was highly familiar to the rat due to previous EEG recordings. Two conventional speakers were placed on the opposite sides outside the cylinder. Auditory stimuli were created through a computer sound card (Sound Blaster 16, Creative Technology Ltd, Singapore, Singapore) and included pure sinusoidal tones of 7, 9 or 11 kHz pitch (tone duration 150 ms, 70 dB, rise/fall time 5 ms). The signal was analog filtered for the 1–1000 Hz band, amplified (× 1000–5000), and digitized at 2 kHz per channel for further processing using a commercial software (Experimenter’s Workbench, DataWave Technologies, Longmont, CO, USA).

At the end of the experiment, the rats were euthanized by an overdose of anesthetic pentobarbital and chloral hydrate each 80 mg/kg i.p. and the sites of the electrode tips were marked by passing a 30 μA anodal current for 5 s through each hippocampal electrode. Subsequently, the brains were immersion-fixed overnight with 4% formalin (formalin was diluted from 37% formaldehyde solution (Sigma-Aldrich)) and sectioned at 50 μm with a vibratome (Leica VT1000s). The sites of the electrolytic lesions were verified in sections stained with cresyl violet Sigma-Aldrich) by using a light Olympus CX microscope Figure 1.

Figure 1. Histological verification of the electrode placement in the hippocampus.

The arrows point to the lesion marks corresponding to the two electrode tips, the upper one in the alveus/oriens and the deeper one in the fissure. Scale bar = 2000 μm.

Study design

The basic study protocol was a conventional mismatch (or oddball) paradigm consisting of one standard tone and one or two deviant tones. Under most conditions, the standard was 9 kHz and the deviants were 7 and 11 kHz tones. Both a low and a high deviant were used to exclude the contribution of tonotopy to auditory evoked potential (AEP) amplitudes. Every run consisted of 400 repetitions with a 1-s inter-stimulus interval. The three tones (7, 9 and 11 kHz) were presented in a pseudo-random order, so that the proportions of the standard, deviant 1 and deviant 2 tones were 85%, 7.5% and 7.5%, respectively.

Experiment 1 consisted of three consecutive days with the 9 kHz tone as the standard, and 7 and 11 kHz tones as the deviants. Similar recordings were performed during Experiment 2 (three weeks after Experiment 1) that also consisted of three consecutive runs. Day 1 replicated Day 1 of the Experiment 1, and was followed by a similar run on Day 2. In addition, Day 2 included a second run with the mismatch contingency reversed, so that 7 kHz became the standard and 9 kHz the deviant. Experiment 3 (one week later) included pharmacological manipulations and consisted only of two runs, one on Day 1 and the second on Day 4. In the first run the standard tone was 9 kHz and the deviants 7 and 11 kHz. In the second run the standard tone was 7 kHz and the deviant 9 kHz. Four rats received scopolamine (0.2 mg/kg, s.c.; Sigma-Aldrich) 20 min before the first run, and five rats before the second run. Saline was used as control treatment.

Data analysis

First, all signals were corrected for amplification. Waveform averaging and AEP peak detection were conducted by custom made routines in Visual Basic under Microsoft Excel^® (version 2002).

The AEP in a typical rat had three middle-latency components, N40, P60 and P110 (N40 means a negative deflection at 40 ms). In addition, these components were followed by a broad negativity from 150 ms to 250 ms after the stimulus onset (Figure 2A, B). The amplitude of these components was calculated as a maximum deviation from the baseline. The baseline was calculated for each rat from the averaged response between 0 and 100 ms before stimulus onset. When calculating mismatch effect between standard and deviant AEP, we focused on the middle-latency components only (N40, P60 and P110).

Figure 2. Representative examples of averaged AEPs obtained in the auditory oddball paradigm.

Cortical (A) and hippocampal (B) AEPs. The thin line denotes the response to the deviant tone and the thick line the response to the standard tone. The triangle marks the tone onset. The horizontal bar corresponds to 100 ms, the vertical bar to 0.04 mV (scale for the cortex is five times smaller than that for the hippocampus). Negativity is downward.

The statistical analysis was conducted by using SPSS for Windows 11.5 software. The standard and deviant responses were compared within-subjects using ANOVA with repeated measures with the run (1–3) or drug (placebo or scopolamine) as additional within-subject factors. The threshold for significance was set to p < 0.05.

Electrode location

Histology verified the location of the hippocampal electrodes in the intended layers: the top electrode in the stratum pyramidale – stratum radiatum and the deeper one in the hippocampal fissure – outer molecular layer of the dentate gyrus. The typical location of the hippocampal electrodes is illustrated in Figure 1.

AEP components

Representative examples of an averaged cortical and hippocampal AEPs obtained in the auditory mismatch paradigm are shown in Figure 2. The components N40, P60 and P110 were identified for each rat and pooled for standards and deviants for all drug-free days. The exact latencies of these components are summarized in Table 1 and their mutual correlations in Table 2. The mutual Pearson correlation coefficients were high and significant for all components of the hippocampal response (if the absolute value of one component grows there is a high probability that other components will also grow). This suggests that physiological source(s) of AEP’s components is not completely independent. On the other hand, only the mutual correlations of the P60–P110 components in the cortical response reached a comparable significance level. Furthermore, neither cortical P60 nor P110 correlated with any hippocampal component, which suggests that the cortical and hippocampal responses are largely independent, with the exception of the early N40 component.

Increased cortical response to the deviant tone

The overall analysis of all three days of Experiment 1 revealed larger cortical responses to the deviant tone compared to the standard tone (Figure 2A, and Figure 3). The difference was significant for N40 [F(1,7) = 7.7, p = 0.03] and P60 (p = 0.04) components and approached significance for P110 (p = 0.06). However, the shape of the average evoked response remained the same, and there was no evidence for the typical mismatch negativity as reported in human studies². In contrast, the hippocampal response did not differentiate between the standard and the deviant tones (p ≥ 0.10 for all components). Together with the correlation table (Table 2) this finding speaks against the notion that the cortical response is a simple volume conducted signal from the hippocampus.

Figure 3. Effect of repetition on the AEP in response to the standard and the deviant tones.

Mean amplitudes of AEP components (N40, P60, P110) ± SEMs are given. In each chart the x-axis represent different runs of the test. Note the break between Run 3 of Experiment 1 and Run 1 of Experiment 2 to indicate the intervening days.

R under Run 3 of Experiment 2 indicates reversal of the mismatch contingency.

* significant difference between consecutive Runs (including between Run 3 of Experiment 1 and Run 1 of Experiment 2);

# significant difference between standard and deviant responses;

& significant repetition effect on component attenuation.

Repetition effect on the responses

The amplitude of cortical N40 response was relatively stable in Experiment 1, but the P60 component attenuated significantly between days [F(2,6) = 5.9, p = 0.04], and the P110 showed a similar, but non-significant trend [F(2,6) = 1.9, p = 0.24]. This trend could be observed for both standard and deviant tones (Figure 3). In contrast, none of the hippocampal components attenuated between days (all p values > 0.40).

The time dependency of AEP attenuation was further investigated in Experiment 2. First, we replicated the standard mismatch condition after 20 intervening days of rest. The cortical response to the standard tones reached the original (or higher) amplitude of Day 1 in Experiment 1 (Figure 3). The ANOVA for repeated measures revealed significant enhancement of cortical P60 [F(1,6) = 12.9, p = 0.01] and P110 (p = 0.03) components between Day 3 of Experiment 1 and Day 1 of Experiment 2. Interestingly, these were the same components that were also attenuated over three daily sessions in Experiment 1. Although a similar trend was observed in the N40 component in some animals, the difference did not reach significance at the group level (p > 0.15). The response enhancement after 20 intervening days could be observed to some extent for both standard and deviant stimulus (Figure 3). In contrast, hippocampal responses, which did not change significantly over the three days of Experiment 1, did not increase after the 20 intervening days of rest, either (all p > 0.35).

Next, we repeated the same mismatch condition on Day 2 of Experiment 2 to see whether this habituation of responses between days could be replicated. This time we saw an attenuation of cortical N40 [F(1,6) = 8.6, p = 0.03] and P60 [F(1,6) = 20.0, p = 0.004] components; and a similar, but not significant trend of P110 component [F(1,6) = 1.7, p = 0.24] (Figure 3). In addition, habituation of hippocampal N40 reached significance [F(1,5) = 12.9, p = 0.02]. Again habituation was similar for the standard and deviant responses. Furthermore, the difference between AEPs to the standard and deviant tones could be replicated. However, this time the most robust oddball effect was observed for cortical P110 [F(1,6) = 29.3, p = 0.002], while P60 showed only a trend (p = 0.07), and N40 no effect (p > 0.30). Unlike in Experiment 1, the hippocampal P60 component showed a clear oddball effect [F(1,5) = 15.2, p = 0.01].

Finally, we reversed the mismatch contingency on the second run of Day 2. The reversal resulted in a robust enhancement of both cortical [F(1,6) = 12.2, p = 0.01] and hippocampal N40 [F(1,5) = 28.7, p = 0.003] components, which increased even above the Day 1 (of Experiment 2) level (Figure 3). This change was observed for both the standard and deviant tones. No other cortical or hippocampal components were enhanced after the reversal (all p > 0.14), but the reversal removed the oddball effect for hippocampal P60 and cortical P110 components (Figure 3).

Scopolamine effect on the middle-latency components

Muscarinic receptors in the central nervous system (CNS) play an important role in the regulation of arousal, attention and synaptic plasticity^16,17. To test the contribution of muscarinic receptors on the mismatch effect, we used the subtype nonspecific muscarinic antagonist, scopolamine¹⁸, in Experiment 3.

Scopolamine resulted in general attenuation of the cortical response, with significant effects in the N40 and P60 components (Figure 4; p = 0.03 and p = 0.04, respectively). In the hippocampal response, only the P60 component decreased significantly (p = 0.002). In Experiment 3, differences were no longer detected between the responses to the standard and deviant sounds for any of the cortical or hippocampal components. Furthermore, the effect of scopolamine did not differ for the standard vs. deviant response (for all sound × drug interactions p > 0.45).

Figure 4. Scopolamine effect on cortical and hippocampus AEPs.

Representative example obtained from one rat in the auditory oddball paradigm. The thin line indicates the response to the deviant tone and the thick line the response to the standard tone. The triangle marks the tone onset. Horizontal bar corresponds to 100 ms, vertical bar to 0.04 mV (cortical scale is 5.7 times smaller than that for the hippocampus). Negativity is downward.

* significant difference between scopolamine and saline runs.