Discrimination of Auditory Stimuli during Isoflurane Anesthesia

Abstract

Deep isoflurane anesthesia initiates a burst suppression pattern in which high-amplitude bursts are preceded by periods of nearly silent electroencephalogram. The burst suppression ratio (BSR) is the percentage of suppression (silent electroencephalogram) during the burst suppression pattern and is one parameter used to assess anesthesia depth. We investigated cortical burst activity in rats in response to different auditory stimuli presented during the burst suppression state. We noted a rapid appearance of bursts and a significant decrease in the BSR during stimulation. The BSR changes were distinctive for the different stimuli applied, and the BSR decreased significantly more when stimulated with a voice familiar to the rat as compared with an unfamiliar voice. These results show that the cortex can show differential sensory responses during deep isoflurane anesthesia.

The electroencephalogram (EEG) shows characteristic temporal patterns under different anesthetic conditions.¹⁷ During deep isoflurane anesthesia, a burst suppression pattern appears in which high-amplitude bursts are followed by periods of nearly silent EEG.^7,11 The burst suppression ratio (BSR) is the percentage of suppression (silent EEG) during the burst suppression pattern and is used to assess anesthesia depth.^11,15 We and others have shown late evoked burst responses synchronized to external stimuli, which might represent unconscious sensory processing.^7,8,17 However a systematic analysis of BSR changes in response to different types of stimuli has not been performed. Because external stimulation during isoflurane anesthesia can initiate cortical burst activity, we investigated whether modality and context-specific cortical activation could occur during this state.

The mechanisms that underlie BSR pattern generation during isoflurane anesthesia are still under investigation. However several key findings have contributed greatly to our understanding. Cortical neural networks show patterns of spontaneous intrinsic activity in the absence of sensory inputs,⁹ and single-neuron membrane potentials in slice preparations can transit spontaneously between 2 states, termed the ‘up’ and ‘down’ states.³ Transitions between these states are abolished by antagonists for glutamate and γ-aminobutyric acid (GABA) receptors.¹⁹ The silent periods of the isoflurane-induced burst suppression activity may be due to long periods of hyperpolarization caused either by inhibition of glutamate-mediated excitatory potentials or by increased GABA receptor-mediated inhibition.^12,22 Isoflurane inhibits voltage-gated sodium channel currents and suppresses the release of glutamate through these blocked channels.²² In addition, the mechanisms that drive suppression periods during isoflurane anesthesia likely involve GABA-gated chloride currents and could also involve increased potassium channel conductance, leading to a reduction of excitatory synaptic input.¹² Evidence also suggests that isoflurane may affect cortical input less than cortical output.⁵

Because specific types of memory can be acquired during isoflurane anesthesia, limited sensory processing may be possible during this state.¹⁰ In addition, certain aspects of cortical information processing seem to be less sensitive to anesthetics than others; therefore, sensory responsiveness to some stimuli may be blocked less effectively by anesthetics at lower doses than other types of sensation, such as painful stimuli.¹ Cortical responsiveness during isoflurane anesthesia may result from a decreased threshold of activation through specific thalamic pathways. Despite the observation that unresponsive conditions are characterized by hyperpolarized thalamic nuclei,¹⁶ sensory stimulation during isoflurane anesthesia can still evoke the appearance of synchronized bursts during burst suppression periods.^7,8 In addition, burst activation may parallel an arousal mechanism, which could remain selectively active during isoflurane anesthesia.²¹

In this experiment, we recorded the EEG continuously in isoflurane-anesthetized Sprague–Dawley rats while applying different types of external auditory stimulation. We hypothesized that external auditory stimulation can elicit modality-specific cortical activation and that some types of sensory discrimination might occur during this state.

Materials and Methods

Eight Sprague-Dawley rats were obtained from Taconic Farms (Germantown, NY) and chronically implanted, under ketamine (100 mg/kg)–xylazine (5 mg/kg) anesthesia, with screw electrodes to record parietal and frontal EEG, and a reference electrode placed over the occipital lobe. Two subcutaneous stainless-steel wire electrodes were placed over the thoracic cavity to record the electrocardiogram, and 2 stainless-steel wire electrodes were inserted into the neck muscles to record the electromyogram (Figure 1). Wires from the recording electrodes were routed to a miniature plug affixed to the skull with dental acrylic that encased the screws and connector. Flunixin meglumine (1.1 mg/kg, subcutaneous) analgesia was given on the day of the surgery and for 2 d afterward. A single intramuscular injection of penicillin G benzathine and penicillin G procaine (10,000 IU/kg) was given to reduce the chance of infection. The rats were maintained on a 12:12-h light:dark cycle (lights on at 07:00) at 22.2 ± 1.1 °C ambient temperature. Water and food were available ad libitum throughout the experiment. Rats within our colony were tested quarterly and certified to be pathogen free by our animal care personal. Surgical and experimental procedures were approved by the Washington State University Animal Care and Use Committee.

Figure 1.

Positions of recording electrodes. Black dots indicate the screw electrodes to record EEG (parietal [PEEG] and frontal [FEEG]); an occipital screw (OEEG) was used as the reference electrode. Neck electrodes for recording of the electromyogram [EMG] are indicated by unfilled circles. B, bregma; L, lambda; T, temporal ridge.

To generate familiarity with voices and enable recording during the transition from wake to anesthesia, rats were movement-restrained by a female trainer and spoken to in daily 2-h sessions starting 3 mo before and continuing for at least 2 mo after the surgical procedures until the animals would remain calm in the restraint for at least 2 h. Movement-restrained training prepared the rats to remain still during these procedures. Restraint also ensured that animals experienced consistent stimuli, because any change in the animal's position can affect the response. The movement-restrained procedure was achieved by enfolding the rats in a piece of soft fabric and then securing them with a denim wrap tightened with hook-and-loop tabs. All animals were trained by the same person daily (Monday through Friday). Although the words used during training were not controlled, they consistently included the same words as were used during the anesthesia studies. Training began when the rats were 2 mo old, with a progressive increase in restraint time by 5 min as the animals remained calm. On average, the rats reached the 2-h mark after 3 mo of training. At this time, surgeries were performed and the training continued for an additional 2 mo. Rats were about 7 mo old when isoflurane experiments were carried out.

During the recording sessions, movement-restrained animals were connected to an EEG amplifier (0.5 to 100 Hz, 256 Hz digitizing rate) and anesthetized by means of a mask delivering 5% isoflurane in oxygen (flow rate, 2 l/min; Fluotec isoflurane vaporizer, Snyder Laboratories, New Philadelphia, OH). Once the EEG showed burst suppression patterns, the anesthetic gas flow rate was adjusted to 0.8 l/min, and 2.3% isoflurane in oxygen (minimum alveolar anesthetic concentration [MAC], 1.5) was applied to maintain the burst suppression pattern and a negative toe pinch response in all animals. By definition, 1 MAC is the dose of an anesthetic gas needed to abolish motor response to a noxious stimulus in 50% of the animals;⁶ 1.5 MAC abolishes the response in >95% of the animals. To achieve consistent burst-suppression patterns with a similar level of surgical anesthesia (negative toe pinch test), we applied 1.5 MAC isoflurane.

The EEG signals were digitized and recorded continuously during both wake and anesthesia by a computer for offline analysis. Electrocardiogram and respiration rate were also monitored, along with measurements of body temperature and CO₂ concentration in expired air. Recordings were performed with and without external auditory stimulation. To measure the effect of different isoflurane anesthesia levels on the BSR, animals initially were recorded for 60 s during each of 4 isoflurane levels: 1.5%, 2.0%, 2.3%, and 3.0%. The BSR then was measured at each level. The baseline 2.3% isoflurane data was also used as a control period during which no external stimulation was applied. The 4 stimulation paradigms were: familiar voice (female trainer) repeating some of the words used during training; unfamiliar (male and female) voices repeating similar words; 0.2-ms single clicks (60 dB intensity; at 1-s regular intervals) generated by an small earphone driven with 5 V; and clapping (approximately 75 dB intensity measured at the animals’ head) at 1-s intervals. A single toe pinch test was applied during each of the different auditory stimulation paradigms to ensure that motor output was negative during the entire test. In addition, to test the effect of a painful stimulus on the BSR, a toe pinch test was applied. The toe pinch test consisted in an intermittent pinch (15 to 20 pinches, 1-s interval, approximately 1 s in duration) applied to a caudal interdigital membrane by using a mosquito hemostat. To produce consistent pinches, the same experimenter performed the toe pinch tests for all animals.

The burst suppression pattern was defined by silent periods in the EEG that were longer than 500 ms, lower than 10 µV, and preceded by high amplitude bursts. The BSR was calculated as the sum of the silent period time divided by the total epoch time (approximately 60 s). Comparison of BSR averages from deep isoflurane anesthesia periods during the presence of different auditory stimuli, painful stimulation (toe pinch), and absence of stimulation were performed by ANOVA (Tukey test, Excel, Microsoft, Bellingham, WA). Differences between each of the conditions and baseline and between conditions were compared by using paired t tests. BSR data are presented as mean and ± SE.

Results

The BSR systematically increased under increasing isoflurane doses (data not shown), as reported in earlier studies.⁸ However, the present study also showed rapid decreases in the BSR during external stimulation under constant isoflurane concentration (2.3%). Cortical burst activity increased during all types of stimulation (Figure 2), such that the BSR varied depending on the type of auditory stimulation applied (Figure 3). The toe pinch stimulation slightly increased burst activity, but the bursts were not synchronized to the stimuli (Figure 2). The bursts occurred more frequently and BSR decreased when the animal heard a familiar voice compared with an unfamiliar voice, but the strongest effect on the BSR was obtained when the auditory stimulation was applied in a form of clapping (Figure 3).

Figure 2.

Discriminative effect of auditory stimulation on cortical burst activity. Examples of the EEG traces recorded during isoflurane anesthesia (2.3%, 1.5 MAC) show representative burst suppression patterns when different auditory stimulation paradigms and the toe pinch test were applied. The numbers on the right are the calculated BSR for each example. Cortical burst activity increased when stimulation was applied, making the BSR significantly lower than in absence of external auditory stimulation. The strongest effect on the BSR was produced by clapping, and familiar and unfamiliar voices generated clearly different effects.

Figure 3.

Effect of external stimulation on cortical burst suppression. BSR values showed significant differences among the different stimulation paradigms (No stimulation, 82.3% ± 2.06%; toe pinch alone, 73.5% ± 4.99%; unfamiliar male voice, 73.4% ± 1.08%; unfamiliar female voice, 69.1% ± 3.03%; earphone click, 67.4% ± 3.54%; familiar female voice, 62.9% ± 2.33%; and clapping, 52.0% ± 4.57%). Pairwise P values from the paired t tests are shown in the table above the bar graph. The familiar voice (female trainer) showed a stronger (P < 0.002) effect on the cortical burst activity than the unfamiliar voice, but the auditory stimulation by means of clapping elicited the highest burst activity. Toe pinch stimulation decreased the BSR slightly but not significantly. These results suggest discrimination of the external stimulation during deep isoflurane anesthesia (1.5 MAC).

Discussion

These results demonstrate that the BSR and the appearance of bursts are influenced by external stimulation. Others have demonstrated burst generation by external stimuli and show a difference in the distribution of burst onsets.⁸ However our systematic evaluation of BSR in response to different stimuli shows that the BSR can be manipulated differentially with external stimulation depending on the type of stimulus and intensity. This finding has important relevance in the operating room if depth of anesthesia is estimated at least in part by using the BSR, especially when paralytics are used.¹³ Our study also reveals context-specific differences in the BSR, indicating that some form of sensory processing and discrimination can occur during isoflurane anesthesia.

Further work is needed to show whether bursts can be generated in a modality-specific manner. For example, if auditory stimulation can generate bursts primarily in the auditory cortex, then selective activation of sensory pathways could support the notion that sensory processing and context discrimination could occur during isoflurane anesthesia. In addition, a previous study⁸ describes differences in waveform and latency of the bursts specific to the modality of stimulation. The thalamus is responsible for transferring somatosensory information to the cerebral cortex and modulates the incoming information. Selective activation of sensory circuits might occur when specific thalamocortical pathways are activated during burst periods. Therefore auditory stimuli may produce selective input to the cortex because cortical activation during burst suppression may not always activate the entire brain. If the stimulus is indeed localized to modality-specific regions, this effect may result in sensory discrimination of particular stimuli. In this study, clapping produced the largest change in BSR, perhaps due to the threatening nature or high intensity of the sound. Even though toe pinch stimulation slightly increased burst activity, the bursts were not synchronized to the stimuli; the painful stimulation may activate nociceptors that fire in a long-lasting manner, longer than the 1-s interstimulus interval used.

Considerable experimental evidence indicates that awake rats can discriminate among human voices,²⁰ and the distinctive changes in the BSR between familiar and unfamiliar voices suggest contextual discrimination of different auditory stimuli under anesthesia. Because the familiar voice was a female trainer and one of the unfamiliar voices was a male experimenter, the animals may have responded to a voice of a particular tone during isoflurane anesthesia. In this case, the same word was uttered with the same loudness but the tone was different. As shown in Figure 3, the burst pattern differed slightly between the unfamiliar male voice and the unfamiliar female voice, perhaps due to the tone, but a familiar female voice tended to elicit even more bursts than the unfamiliar female voices. Although further experiments are needed to confirm the effect of familiarity, these results suggest that discrimination of voices or tone can occur during deep isoflurane anesthesia.

Although our results present a general phenomenon of cortical burst activity in rats, further research is needed to identify the neurophysiologic basis behind the selective initiation of bursts. The invasive nature of this preparation and extensive time needed for the training necessitated the use of a minimal number of experimental animals during this study. Because external stimulation can affect the BSR recorded from the patient, minimal external stimulation that is present during anesthesia induction may cause the BSR to be artificially high, with the result that the patient's anesthetic depth may be less than expected if estimated based in part on the BSR. In addition, because external stimulation can trigger bursts in a stimulation-specific manner, our data support the idea that differential sensory processing may occur during the burst.

The ultimate goal of anesthesia is to eliminate or reduce pain during surgery, both in animals and humans. If anesthesia depth is insufficient, the patient can experience excessive pain and trauma. Excessive anesthesia depth can lead to patient mortality. Assessment of anesthesia depth traditionally includes interpretation of changes in heart rate, blood pressure, respiration, capillary refill rate, BSR, and muscle tone or movement. However, none of these measures can and should be used in isolation. For example, movement may not always reflect anesthesia depth in animals.² Indeed, although an animal may not be able to relate what it recalls or felt during surgery, every practical effort should be made to minimize the potential for painful experiences during surgery. In the human clinical arena, the same principles apply, yet querying patients about their experience is easy to do. Because of this, some surgeries are performed with a variety of amnesic agents to make the patient forget what happened during the surgery.¹⁴ Sometimes paralytics are administered to inhibit movement,¹³ making assessment of anesthetic depth more difficult. Under that condition, the BSR, heart rate, and blood pressure become much more important measures of anesthesia. Awareness reports during anesthesia in humans usually are associated with postsurgical psychologic pathologies,^4,18 and many of those reports might be related to a similar perception of external stimulation during anesthesia, as is seen in the present study.