- Email: [email protected]
Motor, but not sensory, cortical potentials are amplified by high-protein diet
Physiology& Behavior, Vol. 50, pp. 887-893. ©Pergamon Press plc, 1991. Printed in the U.S.A.
0031-9384/91 $3.00 + .00
Motor, But Not Sensory, Cortical Potentials Are Amplified by High-Protein Diet J E F F E R Y W. B R O C K * A N D C H A N D A N P R A S A D * t
*Neuroscience Laboratory, Pennington Biomedical Research Center 6400 Perkins Road, Baton Rouge, LA 70808 "?Section of Endocrinology, Department of Medicine Louisiana State University Medical Center, New Orleans, LA 70112 R e c e i v e d 21 February 1991 BROCK, J. W. AND C. PRASAD. Motor, but not sensory, cortical potentials are amplified by high-protein diet. PHYSIOL BEHAV 50(5) 887-893, 1991.--Animals fed a high-protein diet (50% casein) are hyperactive and more responsive to nociceptive stimuli than those fed either a normal- or low-protein diet. The mechanisms mediating dietary protein-induced behavior are unknown and may include both central and peripheral neural effects. Adult, Sprague-Dawley rats were fed 50% casein (treatment group) and 24% casein (control group) ad lib for 36--40 weeks. The animals were anesthetized with alpha-chloralose and urethane (50 mg/kg and 1.5 mg/kg, IP). EEG recordings were averaged while the anesthetized animal was conditioned using an alerting stimulus-imperative stimulus (AS-IS) paradigm. AS consisted of a 1.5 kHz, 90 dB tone cue. This was followed 2 seconds later by IS, an electrical tail stimulation (11 V, 1.4 s duration). Two negative deflections (N1 and N2) were generated by the frontal cortex during the AS-IS interstimulus interval. N1, an alerting response, was not different between the two groups. N2 amplitude and peak latency were significantly increased in the high-protein group (205% and 117% of control, respectively; p<0.05). N2 represents the activation of cells in the motor cortex. Bralnstem auditory-evoked responses and somatosensory-evoked potentials also were recorded, but no differences were observed between the two diet groups. These data suggest that consumption of a high-protein diet results in an increase in central arousal mechanisms (measured by cortical negativity response), specifically involving increased excitability of the motor cortex, that is not associated with a disorder of information processing in the cerebral cortex (measured by brainstem auditory-evoked responses and somatosensory-evoked potentials). Protein
Evoked potentials
Event-related potentials
Motor cortex
expression of food-searching behavior (22). Another theory is that high dietary protein increases tyrosine availability for the synthesis of central catecholamines which, in turn, increase arousal levels in the animal (6). A weakness of this theory is that plasma and brain tyrosine levels are not elevated in the rat even after chronic consumption of 75% casein, although most neutral amino acids clearly are increased (23). Also, high-protein diets cause ketosis and an increase in the excretion of calcium (15) and magnesium (34), which raise the possibility of peripheral sensory abnormalities, but the contribution of these factors to the observed behavior is uncertain. Examining the neuroelectfical correlates of behavior may distinguish different central or peripheral mechanisms that mediate proteininduced hyperresponsiveness. One of the ways in which the frontal cortex participates in behavioral responses is by receiving auditory, visual, and somatosensory information, then activating or inhibiting specific motor subroutines, in accordance with the demands of the alerting stimuli (17,30). A subject's general state of arousal is markedly expressed in the magnitude of the negative shift in cortical slow potentials when presented with an alerting stimulus that is time coupled to an imperative stimulus. In primates, this type of event-related potential is called the Contingent Negative Variation (CNV) (28). It has been shown that CNV amplitudes are directly related to central dopaminergic activity (35). Rats dem-
RATS fed chronic, high-protein diets (50-80% casein) are easily frightened and demonstrate more defensive behavior than rats on control diets (32). In mice, increasing the protein:carbohydrate ratio in the diet results in increased locomotion in an open field and a more fully developed aggressive response to external stress (33). Such animals are not exceptionally inclined to either dominant or subordinate social behavior (33), nor are they otherwise behaviorally incompetent (22). Previous data from our own laboratory have shown that rats fed a chronic, high-protein diet (50% casein) are hyperactive and are more responsive to nociceptive stimuli than those fed either normal-protein or lowprotein diets (20% and 8% casein, respectively) (22). The behavioral effects of the diet were consistent in both ad lib and pair-fed designs. The mechanisms underlying the protein-induced changes in psychomotor behavior remain unknown. Most studies correlate an increase in locomotor behavior with an increase in the protein:carbohydrate ratio in the diet, but it is often unclear whether increasing protein or decreasing carbohydrate level is the more important variable (22,33). It is equally uncertain if the hyperresponsiveness is the direct effect of a macronutrient or represents an adaptation (22,33). For example, it is known that rats will consume a diet balanced in protein, carbohydrate, and fat in a free choice paradigm (16). When forced to consume a diet of 50% casein, the rats may increase their locomotor activity as an
887
~8~
BROCK ANI-~ P R A S / \ i '
onstrate Cortical Negativity Responses (CNR) that are similar to the CNV in humans, and they are measureable even under urethane/chloralose anesthesia (12,25). If a chronic, high-protein diet stimulates catecholamine activity in the rat, then the animal may be expected to demonstrate an increase in CNR amplitude. Brainstem Auditory Evoked Potentials (BAEP) and Somatosensory Evoked Potentials (SSEP) are measures of auditory and peripheral sensory information processing. These evoked potential tests are sensitive to nutritional (31) and maturational (14) abnormalities. In the present study, CNRs, BAEPs, and SSEPs were recorded in order to investigate the effects of highprotein diet on rat behavior at the neurophysiological level. This experiment was designed to test the hypothesis that consumption of a high-protein diet results in an increase in central arousal mechanisms (measured by CNR) that is associated with a disorder of information processing in the cerebral cortex (measured by BAEPs and SSEPs). The relevance of using rats to study even-related potentials lies in the fact that rats possess an auditory system which is physiologically representative of other mammals (31). Although there have been few previous studies of CNRs in rats, interpretation of these data is strengthened on the basis of the remarkable similarity of the rat CNR to CNVs recorded from primates (25). Hence, they may be used to model human neurobehavioral disorders in a way that is rarely possible. METHOD
Fourteen male, Sprague-Dawley rats were purchased as weanlings from commercial breeders (Harlan Sprague-Dawley). The animals were divided into two groups (7 animals per group). The control group was fed Purina Rat Chow, which contained a normal concentration of protein (23.4% protein; 4.5% fat; 25.3% crude fiber). The treatment group was fed a high-protein (HP) diet (50% protein; 5% fat; 5% crude fiber; 17.2% corn starch; 17.2% sucrose), which was commercially prepared by Research Diets, Incorporated. Both animal groups remained on the diets ad lib for 36-40 weeks, after which they were anesthetized and prepared for experimentation.
Preparation for Surgery On the day of experimentation, the animal was anesthetized using alpha-chloralose and urethane (50 mg/kg and 1.5 mg/kg IP, respectively). After induction of anesthesia, the hair over the animal's skull and along the back over the spine was clipped. The body temperature was maintained at 3 7 - 0 . 5 ° C by a heating pad placed under the animal and monitored by rectal probe. A surgical incision was made ventrally in the neck to expose the trachea. An incision was made in the trachea and PE 240 polyethylene tubing inserted into the airway to facilitate ventilation of the animal. The animal was permitted to breathe spontaneously.
Recording Evoked Potentials Evoked and event-related potentials were recorded using a Nicolet Compact Four Electrodiagnostic System. All records were averaged on-line. Each record was visually inspected for artifact before being stored on disk for off-line analysis.
Brainstem Auditory-Evoked Potentials (BAEPs) After the animal was anesthetized and surgically prepared as described above, its head was fixed in a stereotaxic frame (Stoelting). Ear bar adaptors were positioned in the indention of the animal's squamosal bones. Earplug speakers were inserted
into the left and right auditory meatus and secured m place with surgical tape. Platinum wire electrodes (Grass Instruments) ~tcre inserted subcutaneously over the skull. A reference electrude was positioned over the top of the skull at the midline /Cz!. ~\ctivc electrodes were positioned just behind each ear and lateral m the temporalis muscles (A1 and A2). A ground electrode was positioned over the frontal bone. The auditory stimulus consisted of rarifaction clicks, 100 microseconds in duration, and at a rate ot I 1.4/second. The stimulus was delivered monaurally at different intensity levels (30, 45, 60, 75, and 90 dB nHL). The opposite ear received noisemasking at 15 dB less than stimulus intensity. Two thousand sweeps were averaged, with a sweeptime of 5 milliseconds. The bandpass setting was 150-3 kHz. Two-channel recordings were obtained: Cz-AI, and Cz-A2.
Somatosensor 3' Evoked-Potentials (SS EPs ) A platinum wire electrode was inserted subcutaneously over the frontal bone (Fpz) for reference. Three platinum wire electrodes were used for recording: 1) over the midline of the parietal bone (Cz), 2) subdermally over the seventh lumbar spinous process (spL7), and 3) inserted into the base of the tail (T). A ground electrode was positioned subcutaneously over the interparietal bone. Stimulating electrodes consisting of two stainless steel pins were inserted into the tail of the rat (3 cm apart) and a ground electrode (steel pin) was inserted proximally, near the base of the tail. The stimulus consisted of an electrical shock delivered to the tail with a 200 microsecond duration and at a rate of 4.8/second. Stimulus intensity was adjusted to the minimum amperage required to observe a tail twitch response. Threechannel recordings were obtained; channel 1: Fpz-Cz, channel 2: Fpz-spL7, and channel 3: Fpz-T. For channel 1, the bandpass was 30-1.5 kHz. For channels 2 and 3, the bandpass was 5-1.5 kHz. A minimum of 2000 sweeps were averaged per record. with a sweeptime of 25 milliseconds.
Cortical Negativio' Responses (CNRs) The cranium was surgically exposed. Platinum needle electrodes were placed in contact with the dura mater through small holes made in the skull. A reference electrode was located over the frontal bone (2 mm anterior and 2 mm lateral to bregma). Two active electrodes (A3 and A4) were located over the parietal cortex (7.5 mm posterior, + 4 and - 4 mm lateral to bregma). A ground electrode was placed over the interparietal bone. Two-channel recordings were made (Fpz-A3, Fpz-A4l with a bandpass of 0.01-30 Hz. The alerting stimulus (AS) consisted of a continuous 1.5 kHz tone. The stimulus was delivered at a rate of 3/minute, with a 1 millisecond ramp, a 100 millisecond plateau, and at an intensity of 90 dB nHL. The rats were allowed to habituate, or become familiar with the alerting stimulus, in at least two sessions of exposure, with a minimum of 5 minutes between sessions. Each animal met criteria for habituation when the waveform recording demonstrated no positive deflection in response to AS (a startle response to the alerting stimulus). The recording electrodes were strategically positioned proximal to the auditory cortex (37) in order to facilitate recording of the startle response. This permitted a reliable determination of habituation without compromising the recording of expectancy negativity, which has both a frontal and parietal distribution (4,13). Eighty trials were averaged during each session, with a sweeptime of 5 seconds. All sweeps were externally triggered 1 second prior to AS. After the habituation sessions, AS was paired with the imperative stimulus (IS). To accomplish this, stimulating electrodes consisting of two stainless steel pins were inserted into the tail of the rat (3 cm apart). A ground electrode (steel pin) was inserted more proximally, near the base of the tail. IS consisted of an 11-volt, square wave pulse of 1.4
CORTICAL RESPONSES IN HIGH-PROTEIN DIET
889
second duration (Grass Stimulator, Model $48; Grass Stimulus Isolation Unit, Model SIU5B). The animal was trained with AS-IS in 3 sessions, with interstimulus intervals (ISI) of 0.5, 1.0, and 2.0 seconds, respectively. The intensity of the tail shock was kept constant for each session. Again, 80 trials were averaged in each session and intertrial intervals were at least 5 minutes apart. Following training to the AS-IS paradigm with 2-second ISI, another session was obtained in which IS was withheld, allowing the animal to undergo extinction of training. In a final session, the animal was retrained to AS-IS with a 2-second ISI.
I 1pV HABITUATION
+
EXTINCTION
Data Analysis RETRAINING
The latency of each component was measured from the onset of the stimulus to each deflection's highest point. Amplitude was calculated as the voltage difference measured peak-to-trough. The latencies and amplitudes of the peaks of the two diet groups were analyzed using two-tailed, unpaired Student's t-test. In all analyses, statistical significance was assumed when p < 0 . 0 5 was obtained.
AS I 0
I 1
t I z
I 3
I 4
I 5
SECONDS
RESULTS
Body weights were not different between the two groups at the time of sacrifice (Controls: 502---14 grams; HP group: 495--- 16 grams). On the day of experimentation, animals consuming the HP diet were observed to be more alarmed and resistive to handling than the counterparts on a normal diet, although these observations were not quantitated. Both groups, however, achieved an adequate level of anesthesia with the same dosages of urethane and alpha-chloralose. Figure 1 shows the results recorded from one animal during the CNR protocol. In the HP group, all animals habituated to the alerting stimulus within two sessions. Within the control group, 4 of the 7 animals required a third habituation session in order to meet criteria for habituation. When the imperative stimulus was paired with the alerting stimulus, all animals generated two distinct negative deflections during the interstimulus interval. Figure 2 shows the grand average CNR for one animal in the study. The first negative deflection (N1) was characterized by a sharp rise, early peak, and short duration. The second negative peak (N2) was characterized by a later peak, a slower rise time, and a much longer duration. N2 was occasionally bifurcated. The typical response to the imperative stimulus was a gradual, positive baseline shift in the waveform. After training with AS-IS with a 2-second interstimulus interval, the animals were allowed to undergo extinction reactions by discontinuing the imperative stimulus. Then each animal was retrained with AS coupled to IS, recreating the CNR. Absolute latencies, voltages, and durations for N1 and N2 are presented in Table 1. All data for Table 1 were calculated from the grand average waveforms, such as the one depicted in Fig. 2. N1 latencies, amplitudes, and duration were not statistically different between the two diet groups. There was a tendency for N1 amplitudes to be larger in the HP diet group, but the differences were not significant at the 95% confidence level. Regarding the N2 deflection, animals consuming the HP diet demonstrated significantly longer peak latencies (p<0.05). The HP group also demonstrated significantly larger N2 amplitudes (p<0.05). N2 durations were not different between the two groups. BAEPs were successfully recorded in all animals. Four positive deflections were reproducible. These were identified as Waves I through IV, typically seen in rodent short-latency auditory-evoked responses. Table 2 shows the absolute peak laten-
IS
t
FIG. 1. Cortical negativity responses (CNR) recorded from one urethane/ chloralose anesthetized rat in this study. The animal was conditioned using an alerting stimulus-imperative stimulus paradigm. Each EEG tracing is the average of 80 sweeps. In the top tracing, the animal was habituated to the alerting stimulus (AS), which was a tone cue. When the tone cue was followed by the imperative stimulus (IS), which was an electrical tail shock, there was a negative shift in the EEG baseline during the interstimulus interval. Withholding the IS resulted in extinction of the negative shift response. In the bottom tracing, the A 1 was coupled again to the IS, recreating the CNR.
cies of Wave IV recorded from the two diet groups, as well as the I-IV interpeak intervals. As can be seen, there were no differences between the two diet groups at any intensity level tested. Figure 3 shows the audiometric analyses of the animals in both diet groups. No reproducible peaks occurred when the animals were stimulated at 30 dB. Wave IV latencies were max-
I lpV +
N1
AS
IS
t I 0
I 1
NZ
t I Z
I 3
I 4
I 5
$ECONOS FIG. 2. Grand average (160 sweeps) of the two CNR responses depicted in Fig. 1. This technique improves the signal-to-noise ratio and more clearly defines two negative deflections (N1 and N2) generated by the frontal cortex when the animal has been conditioned to the alerting stimulus (AS) followed by an imperative stimulus (IS).
;',;90
B ROCK A N l) P R -\ S~,:\i
TABLE I PEAK LATENC1ES AND AMPLITUDES OF CNR PEAKS C o n t r o l Diet
N1 Latency(ms) (mean -+ SEM) Amplitude (IJ,V) (mean ± SEM) Duration (ms) (mean m SEM)
931 ± 93
High-Protein Diet
N2
NI
N2
1330 _+_ 1
1120 _+ 52
1560 ± 1"
0.75 ± 0.11
1.11 ± 0.13
1.74 ± 0.57
2.28 ± 0.25*
277 +_ 81
684 ± 91
244 ± 55
962 _+ 196
*Indicates p<0.05, control diet (N = 7) versus high-protein diet (N = 7), unpaired Student's t, two-tailed test. The data were measured from CNR grand averages for each animal, as depicted in Fig. 2.
imal at intensity levels at or above 75 dB. Also, the amplitudes of the different peaks were not different between the two groups. At 45 dB, Wave IV amplitudes averaged 0.04-+0.01 ixV for both diet groups. At 60 dB, Wave IV amplitudes in the control group were 0 . 1 3 ± 0 . 0 4 txV, versus 0 . 0 9 ± 0 . 0 2 txV in the HP group. At 75 dB, Wave IV amplitudes plateaued at 0.21 ± 0 . 0 6 I~V in the control group and 0 . 1 8 ± 0 . 0 3 txV in the HP diet group. Also, at 90 dB, I/IV amplitude ratios were not different between the two groups (control diet group: 4.47 -+ 1.02; HP diet group: 6.76-+1.69). In total, there were no significant differences in the BEAPs between the two diet groups. Somatosensory-evoked responses were also successfully recorded in all animals. Stimulus intensities required to observe a tail twitch response in the control diet group were 4.2 ± 0.9 mA, while the HP diet group required tail stimulus intensities of 3.7---0.6 mA. Tail stimulation resulted in 3 reproducible deflections in the waveforms, designated according to their peak latencies as N1.8, N5, and P16. In all animals, a negative deflection was most predominant in the Fpz-T channel, representing depolarization of the peripheral nerves in the tail. This deflection occurred with a peak latency of 1.86-+0.14 ms in the control animals, and 1.63 -+ 0.14 ms in the HP diet animals (not signifi-
4.500 O m
4.000 lip •
3.500
~
3.000 2.500
30
!
i
i
45
60
75
90
Click I n t e n s i t y (dB nHL) FIG. 3. Audiometric analyses of brainstem auditory evoked potentials recorded from animals fed the medium-protein diet (C, N = 7) and highprotein diet (HP, N = 7) (mean-+-SE). Wave IV response latency is plotted as a function of click intensity. Threshold for Wave IV generation was between 30-45 dB nHL. The slope of the latency-intensity function describes normal sensoryneural recruitment and information conduction.
cant). A second negative deflection was most easily identified in the Fpz-spL7 channel, and represented depolarization of the spinal cord, as the peripheral nerve enters the spinal cord and synapses. In the control animals, this negativity occurred with a peak latency of 5.44__.0.24 ms, and in the HP diet animals at 5.33-+0.10 ms (not significant). Peripheral conduction times were determined for the two groups by calculating the N1.8-N5 interpeak intervals, and were found to be not significantly different (control diet group: 3.58-+0.21 ms; HP diet group: 3.70-+0.15 ms). A positive deflection (P16) was identifiable in all 3 channels (all referenced to Fpz) which represented the response of the sensory cortex to the peripheral nerve stimulation. The absolute latencies and amplitudes of the PI6 responses recorded from the two diet groups are shown in Table 3, as well as calculated central conduction times (N5-P16 interpeak intervals). In total, there were no significant differences in the SSEPs between the two diet groups. DISCUSSION
The major observation made from this study was that animals fed a chronic, high-protein diet demonstrated CNRs which where significantly larger in amplitude and longer in latency than CNRs recorded from control animals (Table 1). The difference in CNR amplitude stands in contrast to the other recorded variables, BAEPs and SSEPs (Tables 2 and 3, respectively), which were not different between the two groups. It is important to note that differences between the diet groups are not explainable based upon differing levels of anesthesia. Although the animals consuming the HP diet acted more alarmed and resistive to handling than the control animals, both achieved an adequate level of anesthesia with the same dosages of urethane and alpha-chloralose. With the dosages used in this study, alpha-chloralose behaves as a sedative-hypnotic agent, while the urethane induces a profound and very long-lasting condition of anesthesia in the animal. The plane of anesthesia in each animal was monitored by toe clamp and whisker reflexes throughout the experiment. Urethane (ethyl carbamate) has often been the anesthetic of choice for neuroscientists, because it has less influence than other anesthetics on the normal activity of a wide variety of brain areas, including the cerebral cortex ( I l L In evoked potential studies, urethane anesthesia typically reduces amplitudes and increases latencies, but does not produce fundamental changes in wave morphology (2). Neither does the drug disrupt cortical mechanisms involved in learning, even at anesthetic dosages (25,26). Although the mechanism of action of urethane is unknown (11), the remarkable features of this drug
CORTICAL RESPONSES IN HIGH-PROTEIN DIET
891
TABLE 2 LATENCIES,AMPLITUDES,AND INTERPEAKINTERVALSOF BEAP PEAKS* Control Diet
Latency (ms) (mean - SEM) Amplitude (I~V) (mean --- SEM)
High-Protein Diet
Wave IV
I-IV Interval
Wave IV
I-IV Interval
2.94 _ 0.15
2.55 --- 0.13
2.86 --- 0.06
2.43 --- 0.05
0.24 --- 0.06
0.19 --- 0.04
*Recorded at 90 dB nHL. All data: p>0.05, unpaired Student's t, two-tailed test.
may be exploited to perform classical conditioning in a fully anesthetized animal. Wave I of the brainstem auditory potentials arises from depolarization of the auditory nerve; wave IV arises from the vicinity of the ipsilateral superior olive, and possibly as proximal as the lateral lemniscus; the I-IV interpeak interval is a measure of brainstem transmission time for the auditory information (31). Normal BEAPs indicate that the high-protein diet did not alter auditory information processing, at least at the level of the bralnstem. Similarly, normal SSEPs indicate that the high-protein diet had no obvious effect on the peripheral response to electrical stimulation of the tall. The P16 component is similar to the positive cortical potentials observed by others (14). This component is apparently generated by postsynaptic activation of cell bodies located in the cortex in response to thalamic input (1). The data suggests that the abnormal CNRs induced by the HP diet are not explained by changes in peripheral sensory mechanisms, such as alterations in sensory thresholds or defects in auditory acuity (24). The N1 component which preceded the CNRs in the two diet groups was not different (Table 1). This component is believed to be generated by the frontal cortex as an orienting response to the alerting stimulus (10) and reflects the processing of directional information (4). In primates, the N1 is not necessarily related to the contingency of the CNV paradigm (10). However, in the present study, generation of the frontal component was contingent with the imperative stimulus (Fig. 1). The rat frontal cortex responded to the AS-IS paradigm by generating a waveform that resembled primate frontal cortical potentials, but also demonstrated a contingency which resembled premotor cortical potentials. This observation is consistent with evidence that the
frontal cortex of the rat is not homologous to that of the primate, but is more similar anatomically to the primate cingulate and premotor cortices (17). Normal N1 responses reinforce the conclusion that auditory information processing in the HP diet group was not different from the control group. As mentioned previously, the CNR in the rat is analogous to the CNV recorded in primates, which strengthens our ability to understand the implications of these data. There is no doubt that the CNV amplitude depends upon the learning of the relevance of the AS-IS contingency (28). The precentral motor area is responsible for generation of this waveform and its appearance represents the preparation or execution of motor action by the subject (29). One important aspect of motor preparation is the condition of "motor set," which refers to the state of readiness of the cells of the motor cortex to generate a command for movement (38). In theory, cortical cell activity is determined at any moment by a previously learned motor program. The motor program may be changed on the basis of new sensory information and result in a different level of cell activity (38). Within the framework of this theory, the associative properties of AS processing changed the motor set of the rat cortical cells. Under the condition of the HP diet, amplification of the CNR may.be explained by an abnormality in the cellular process of motor programming that biased a larger area of cells toward synchronous depolarization. The data may also be explained by prolonged activation of the neural circuits involved in analyzing the associative properties of the alerting stimulus. The increase in CNR amplitude in the HP group was accompanied by an increase in absolute peak latency (Table 1). An increase in peak latency was unexpected, since rats on the HP diet demonstrated decreased response latencies to tall flick stim-
TABLE 3 LATENCIES,AMPLITUDES,AND INTERPEAKINTERVALSOF SSEP PEAKS* Control Diet N5-P16 P 16 Interval Latency (ms) (mean +__ SEM) Amplitude (I~V) (mean --- SEM)
16.03 ± 0.82 0.10 --- 0.02
10.58 -+ 0.82
High-Protein Diet N5-P16 P16 Interval 15.98 ± 0.56
10.68 ± 0.61
0.08 --- 0.01
*Recorded at 4.17 --- 0.95 mA (Control diet, N=7), 3.67 _+ 0.67 mA (HP diet, N=7). All data: p>0.05, unpaired Student's t, two-tailed test.
,~2
ulation (22). It was expected that motor response latency would be directly related to CNV peak latency. One possible explanation for this discrepancy is that CNV peak latency does not correlate to motor responses of the purely reflexive type. such as those elicited by the tail flick method. This observation underscores the fact that the CNV is uniquely related to motoric activity that involves the frontal cortex. Consider also that the CNV is a composite of negative deflections related to anticipation, cognition, and motor command, which may be evoked independently or in concert, according to experimental design (13). Measuring the latency of a single deflection in the CNV when a bifid wave is present may not be informative unless we are certain which feature of AS processing it represents. These data are in support of the theory that a chronic, highprotein diet results in an increase in preparatory arousal mechanisms. Tecce conceptualized "'arousal" as a hypothetical process which energizes behavior unselectively, affecting only intensity of response and being devoid of steering properties (28). Although there are cells in the premotor cortex which show directional specificity (38), the amplitudes of premotor potentials appear to correlate more strongly with the force of movement rather than the direction (4). This study sheds more light upon our earlier observations that rats fed a chronic, high-protein diet demonstrate behavioral hyperactivity and hyperresponsiveness to noxious stimuli (22). The present study would suggest that an alteration in central mechanisms is responsible for protein-induced hyperactivity and hyperresponsiveness. Also, the highamplitude CNRs observed in the HP diet group most likely represent an overexpression of motor cortical activity which is generated endogenously, not reflexively, since abnormalities in peripheral sensory mechanisms made no significant contribution to the observed behavior. These data also provide additional support for the suggestion that it is the protein content of the diet which correlated with the observed behavior, and not the carbohydrate content. Since carbohydrate diets increase tryptophan levels in the brain (33), which is known to decrease locomotor activity in rats (33,36), we would predict the carbohydrate effect on CNR amplitudes to be just the opposite of what was seen in this study. Therefore, it is unlikely that the carbohydrate content of the HP diet used in this study contributed to the high amplitude CNRs. Although dopamine was not measured in this study, the data are consistent with the theory that a high-protein diet stimulates dopaminergic activity, as suggested by the high amplitude CNRs (35). At this point, it is interesting that 4 of the 7 control animals required a third habituation session, whereas the HP animals habituated within two sessions. Although the experiments were not designed to precisely quantitate the onset of habituation, it is possible that the HP animals habituated more quickly than the control animals. Habituation normally involves the process of selective attention (7), and both are disrupted by dopaminergic stimulation (7,20). If the HP diet was stimulating a dopaminergic mechanism, then it should have taken longer for the HP animals to habituate. In the anesthetized animal which may not demonstrate selective attention, habituation may result from lower brain functions involved in attentional maintenance (9). Maintenance of attention is increased by L-DOPA (8) and is more directly correlated with basal arousal levels (28). The HP diet may have resulted in an enhancement of attentional maintenance (via increased basal arousal level) in such a way that preserved the anesthetized animal's ability to perform feature analysis of repetitive stimuli. This may have compensated for a delay in habituation otherwise seen in the anesthetized control animals. In humans, overexpression of dopaminergic mechanisms is
BROCK AN[) PRASAi).
involved in schizophrenia, a condition that is well characterized by low amplitude and prolonged CNVs ~35). In the presem study, animals fed a high-protein diet indeed had prolonge~ CNRs, but generated robust, rather than low-amplitude responses. Schizophrenics have normal BAEPs (24) but prolonged cortical SSEPs (3), both of which are normal in the high dietary protein group. Furthermore, schizophrenics more often present with either elevated or normal serum magnesium levels (5}. The electrophysiological effects of a high-casein diet presented herein do not fit very well with a model of schizophrenia. Also, the CNRs recorded from the HP diet group do not compare well to CNVs recorded from hyperactive humans. Such people present with a short latency, but rapidly deteriorating, frontal component in their CNVs, concomitant with an attention deficit, perhaps due to an impaired communication between the frontal cortex and thalamus (35). The frontal component in both diet groups demonstrated high variability, but the HP diet group NI showed a greater tendency to be prolonged compared to controls. High-amplitude CNVs similar to the CNRs observed in the high-protein group are seen in neurotic patients and in cases of psychosomatic illnesses, such as asthma (35). This raises the possibility that a dietary factor may be involved in these illnesses. A very important aspect of the rat CNR which must be mentioned is the fact that there is a cholinergic component, as well as a dopaminergic one, involved in the magnitude of the CNR (12,25). The frontal cortex in the rat receives its cholinergic innervation from the magnocellular forebrain nuclei (27). Presently, it is not clear what the relative roles are for the two neurotransmitter systems in the generation of eventcelated potentials. Others have shown that dopamine D 2 receptor activation has a facilitory effect on cholinergic depolarization of frontal cortical cells (39). The implication is that a high-protein diet causes an increase in dopaminergic activity which may have facilitated cholinergic activity in the frontal cortex, thus amplifying the negativity response. The cholinergic component of the CNR is modulated by the activity of gamma-aminobutyric acid (GABA), as shown by the fact that the negativity response in rats is partially inhibited by the microinjection of GABA into the magnocellular forebrain area (27). Investigators currently believe that diminished GABAergic activity in the cerebrum is an important mechanism involved in anxiety (21). Therefore, CNR recording in rats may have valuable applications in the study of panic disorder or other anxiety-related illnesses (18). Cortical negativity responses are only a component in the expression of motoric behavior, in which other brain regions are involved. In the present study, as in most CNV measurements, generation of the negativity response is constrained by the AS-IS paradigm. This cortical activity may be related to voluntary acts of the Process 1 type (19), which includes CNVs that are paced, or generated in association with sensory cues. With the present limits of knowledge, these must be held distinct from cortical activity of the Process II type, i.e., CNVs that are uniquely associated with fully independent, volitional acts (19). ACKNOWLEDGEMENTS This work was supported by the US Army Research and Development Command, grant DAMD-17-88-Z-8023. Opinions, interpretations, conclusions, and recommendations are those of the author and are not necessarily endorsed by the US Army. In conducting research using animals, the investigators adhered to the Guide for the Care and Use of Laboratory Animals, prepared by the Committee on Care and Use of Laboratory Animals of the Institute of Laboratory Animal Resources, National Research Council (NIH), publication #86-23, revised 1985.
CORTICAL RESPONSES IN HIGH-PROTEIN DIET
893
REFERENCES 1. Addae, J. I.; Stone, T. W. Purine receptors and kynnrenic acid modulate the somatosensory evoked potential in rat cerebral cortex. Electroencephalogr. Clin. Neurophysiol. 69:186-189; 1988. 2. Angel, A.; Gratton, D. A. The effect of anaesthetic agents on cerebral cortical responses in the rat. Br. J. Pharmacol. 76:541-549; 1982. 3. Boning, J.; Drechsler, F.; Neuhauser, B. Somatosensory event-related potentials and selective attention impairment in young chronic schizophrenics. Neuropsychobiology 21:146--151; 1989. 4. Bonnet, M.; Macday, W. A. Changes in contingent-negative variation and reaction time related to precueing of direction and force of a forearm movement. Brain Behav. Evol. 33:147-152; 1989. 5. Brackenridge, C. J.; Jones, I. H. Relation of hypermagnesemia to activity and neuroleptic drug therapy in schizophrenic states. J. Neurol. Neurosurg. Psychiatry 34:195-199; 1971. 6. Chafetz, M. D. Catecholamines and behavioral arousal. In: Chafetz, M. D., ed. Nutrition and neurotransmitters, the nutrient bases of behavior. Englewood Cliffs, NJ: Prentice-Hall, Inc.; 1990:67-87. 7. Cheal, M. L. Disruption of selective attention by apomorphine, but not amphetamine, in the Mongolian gerbil. Psychopharmacology (Berlin) 69:93-100; 1980. 8. Cheal, M. L. L-DOPA and piribedil alter different components of attentional behavior dependent on dose. Psychopharmacology (Berlin) 80:345-350; 1983. 9. Cheal, M. Attention and habituation: Separation of specific components in development. Physiol. Behav. 41:415-425; 1987. 10. Desmedt, J. E.; Debecker, J. Waveform and neural mechanism of the decision P350 elicited without pre-stimulus CNV or readiness potential in random sequences of near-threshold auditory clicks and finger stimuli. Electroencephalogr. Clin. Nenrophysiol. 47:648-670; 1979. ll. Dyer, R. S.; Rigdon, G. C. Urethane affects the rat visual system subanesthetic doses. Physiol. Behav. 41:327-330; 1987. 12. Ebenezer, I. Cortical slow potentials recorded in rats anaesthetized with urethane are affected by nicotine. Br. J. Pharmacol. 79:270P; 1983. 13. Frost, B. G.; Neill, R. A.; Fenelon, B. The determinants of the non-motoric CNV in a complex variable foreperiod, information processing paradigm. Biol. Psychol. 27:1-21; 1988. 14. Gilmore, R.; Brock, J.; Hermansen, M. C.; Baumann, R. Development of lumbar spinal cord and cortical evoked potentials after tibial nerve stimulation in the pre-term newborns: Effects of gestational age and other factors. Electroencephalogr. Clin. Neurophysiol. 68: 28-39; 1987. 15. Holl, M. G.; Allen, L. H. Comparative effects of meals high in protein, sucrose, or starch on human mineral metabolism and insulin secretion. Am. J. Clin. Nutr. 48:1219-1225; 1988. 16. Kanarek, R. B. Determinants of dietary self-selection in experimental animals. Am. J. Clin. Nutr. 42:940-950; 1985. 17. Kolb, B. Functions of the frontal cortex of the rat: A comparative review. Brain Res. Rev. 8:65-98; 1984. 18. Lader, M.; Mathews, A. Physiological changes during spontaneous panic attacks. J. Psychosom. Res. 14(4):377-382; 1970. 19. Libet, B.; Wright, E. W.; Gleason, C. A. Readiness-potentials preceding unrestricted "spontaneous" vs. pre-planned voluntary acts. Electroencephalogr. Clin. Neurophysiol. 54:322-335; 1982. 20. Matthysse, S. Dopamine and selective attention. In: Costa, E.; Gessa, G. L., eds. Advances in biochemical pharmacology, vol. 16. New York: Raven Press; 1977:667--669.
21. Nestoros, J. N. Gabaergic mechanisms and anxiety: An overview and a new neurophysiological model. Can. J. Psychol. 29:520-529; 1984. 22. Onaivi, E. S.; Brock, J. W.; Prasad, C. Behavior modification in rats involving changes in dietary protein. Nutr. Res., submitted; 1991. 23. Peters, J. C.; Harper, A. E. Adaptation of rats to diets containing different levels of protein: Effects on food intake, plasma and brain amino acid concentrations and brain neurotransmitter metabolism. J. Nutr. 115:382-398; 1985. 24. Pfefferbaum, A.; Horvath, T. B.; Roth, W. T.; Tinklenberg, J. R.; Kopell, B. S. Auditory brain stem and cortical evoked potentials in schizophrenia. Biol. Psychol. 15(2):209-223; 1980. 25. Pirch, J. H.; Corbus, M. J.; Rigdon, G. R. Event-related slow potentials in urethane anesthetized rats. Soc. Neurosci. Abstr. 9:1197; 1983. 26. Pirch, J. H.; Corbus, M. J.; Ebenezer, I. Conditioned cortical slow potential responses in urethane anesthetized rats. Int. J. Neurosci. 25:207-218; 1985. 27. Ridgon, G. C.; Pirch, J. H. Microinjection of procaine or GABA into the nucleus basalis magnocellularis affects cue-elicited unit responses in the rat frontal cortex. Exp. Neurol. 85:283-296; 1984. 28. Rockstroh, B.; Elbert, T.; Birbaumer, N.; Lutzenberger, W. Slow brain potentials and behavior. Baltimore, MD: Urban and Schwarzenberg; 1982. 29. Rohrbaugh, J. W.; Syndulko, K.; Sanquist, T. F.; Lindsley, D. B. Synthesis of the contingent negative variation brain potential from noncontingent stimulus and motor elements. Science 208:11651168; 1980. 30. Rosenkilde, C. E. Functions of the prefrontal cortex: Behavioral investigations using ablation and electrophysiological techniques in rats, cats, dogs and monkeys. Acta Physiol. Scand. (Suppl. 514):658; 1983. 31. Shaw, N. A. The auditory evoked potential in the rat--a review. Prog. Neurobiol. 31:19--45; 1988. 32. Sos, J.; Szelenyi, I. Diets for animal experiments. Budapest: Academiai Kiado; 1974. 33. Spring, B. Effects of foods and nutrients on the behavior of normal individuals. In: Wurtman, R. J.; Wurtman, J. J., eds. Nutrition and the brain, vol. 7. New York: Raven Press; 1986:1-47. 34. Stallings, V. A.; Archibald, E. H.; Pencharz, P. B. Potassium, magnesium, and calcium balance in obese adolescents on a proteinsparing modified fast. Am. J. Clin. Nutr. 47:220-224; 1988. 35. Timsit-Berthier, M. Contingent negative variation (CNV) in psychiatry. In: MeCallum, W. C.; Zappoli, R.; Denoth, F., eds. Cerebral psychophysiology: Studies in event-related potentials. EEG (Suppl. 3):429-434; 1986. 36. Warbritton, J. D.; Stewart, R. M.; Baldessarini, R. J. Decreased locomotor activity and attenuation of amphetamine hyperactivity with intraventricular infusion of serotonin in the rat. Brain Res. 143:373-382; 1978. 37. Webster, W. R. Auditory system. In: Paxinos, G., ed. The rat nervous system, vol. 2. Sydney, Australia: Academic Press; 1985:153184. 38. Wise, S. P. The primate premotor cortex: Past, present, and preparatory. Annu. Rev. Neurosci. 8:1-19; 1985. 39. Yang, C. R.; Mogenson, G. J. Dopaminergic modulation of cholinergic responses in rat medial prefrontal cortex: An electrophysiological study. Brain Res. 542:271-281; 1990.
0031-9384/91 $3.00 + .00
Motor, But Not Sensory, Cortical Potentials Are Amplified by High-Protein Diet J E F F E R Y W. B R O C K * A N D C H A N D A N P R A S A D * t
*Neuroscience Laboratory, Pennington Biomedical Research Center 6400 Perkins Road, Baton Rouge, LA 70808 "?Section of Endocrinology, Department of Medicine Louisiana State University Medical Center, New Orleans, LA 70112 R e c e i v e d 21 February 1991 BROCK, J. W. AND C. PRASAD. Motor, but not sensory, cortical potentials are amplified by high-protein diet. PHYSIOL BEHAV 50(5) 887-893, 1991.--Animals fed a high-protein diet (50% casein) are hyperactive and more responsive to nociceptive stimuli than those fed either a normal- or low-protein diet. The mechanisms mediating dietary protein-induced behavior are unknown and may include both central and peripheral neural effects. Adult, Sprague-Dawley rats were fed 50% casein (treatment group) and 24% casein (control group) ad lib for 36--40 weeks. The animals were anesthetized with alpha-chloralose and urethane (50 mg/kg and 1.5 mg/kg, IP). EEG recordings were averaged while the anesthetized animal was conditioned using an alerting stimulus-imperative stimulus (AS-IS) paradigm. AS consisted of a 1.5 kHz, 90 dB tone cue. This was followed 2 seconds later by IS, an electrical tail stimulation (11 V, 1.4 s duration). Two negative deflections (N1 and N2) were generated by the frontal cortex during the AS-IS interstimulus interval. N1, an alerting response, was not different between the two groups. N2 amplitude and peak latency were significantly increased in the high-protein group (205% and 117% of control, respectively; p<0.05). N2 represents the activation of cells in the motor cortex. Bralnstem auditory-evoked responses and somatosensory-evoked potentials also were recorded, but no differences were observed between the two diet groups. These data suggest that consumption of a high-protein diet results in an increase in central arousal mechanisms (measured by cortical negativity response), specifically involving increased excitability of the motor cortex, that is not associated with a disorder of information processing in the cerebral cortex (measured by brainstem auditory-evoked responses and somatosensory-evoked potentials). Protein
Evoked potentials
Event-related potentials
Motor cortex
expression of food-searching behavior (22). Another theory is that high dietary protein increases tyrosine availability for the synthesis of central catecholamines which, in turn, increase arousal levels in the animal (6). A weakness of this theory is that plasma and brain tyrosine levels are not elevated in the rat even after chronic consumption of 75% casein, although most neutral amino acids clearly are increased (23). Also, high-protein diets cause ketosis and an increase in the excretion of calcium (15) and magnesium (34), which raise the possibility of peripheral sensory abnormalities, but the contribution of these factors to the observed behavior is uncertain. Examining the neuroelectfical correlates of behavior may distinguish different central or peripheral mechanisms that mediate proteininduced hyperresponsiveness. One of the ways in which the frontal cortex participates in behavioral responses is by receiving auditory, visual, and somatosensory information, then activating or inhibiting specific motor subroutines, in accordance with the demands of the alerting stimuli (17,30). A subject's general state of arousal is markedly expressed in the magnitude of the negative shift in cortical slow potentials when presented with an alerting stimulus that is time coupled to an imperative stimulus. In primates, this type of event-related potential is called the Contingent Negative Variation (CNV) (28). It has been shown that CNV amplitudes are directly related to central dopaminergic activity (35). Rats dem-
RATS fed chronic, high-protein diets (50-80% casein) are easily frightened and demonstrate more defensive behavior than rats on control diets (32). In mice, increasing the protein:carbohydrate ratio in the diet results in increased locomotion in an open field and a more fully developed aggressive response to external stress (33). Such animals are not exceptionally inclined to either dominant or subordinate social behavior (33), nor are they otherwise behaviorally incompetent (22). Previous data from our own laboratory have shown that rats fed a chronic, high-protein diet (50% casein) are hyperactive and are more responsive to nociceptive stimuli than those fed either normal-protein or lowprotein diets (20% and 8% casein, respectively) (22). The behavioral effects of the diet were consistent in both ad lib and pair-fed designs. The mechanisms underlying the protein-induced changes in psychomotor behavior remain unknown. Most studies correlate an increase in locomotor behavior with an increase in the protein:carbohydrate ratio in the diet, but it is often unclear whether increasing protein or decreasing carbohydrate level is the more important variable (22,33). It is equally uncertain if the hyperresponsiveness is the direct effect of a macronutrient or represents an adaptation (22,33). For example, it is known that rats will consume a diet balanced in protein, carbohydrate, and fat in a free choice paradigm (16). When forced to consume a diet of 50% casein, the rats may increase their locomotor activity as an
887
~8~
BROCK ANI-~ P R A S / \ i '
onstrate Cortical Negativity Responses (CNR) that are similar to the CNV in humans, and they are measureable even under urethane/chloralose anesthesia (12,25). If a chronic, high-protein diet stimulates catecholamine activity in the rat, then the animal may be expected to demonstrate an increase in CNR amplitude. Brainstem Auditory Evoked Potentials (BAEP) and Somatosensory Evoked Potentials (SSEP) are measures of auditory and peripheral sensory information processing. These evoked potential tests are sensitive to nutritional (31) and maturational (14) abnormalities. In the present study, CNRs, BAEPs, and SSEPs were recorded in order to investigate the effects of highprotein diet on rat behavior at the neurophysiological level. This experiment was designed to test the hypothesis that consumption of a high-protein diet results in an increase in central arousal mechanisms (measured by CNR) that is associated with a disorder of information processing in the cerebral cortex (measured by BAEPs and SSEPs). The relevance of using rats to study even-related potentials lies in the fact that rats possess an auditory system which is physiologically representative of other mammals (31). Although there have been few previous studies of CNRs in rats, interpretation of these data is strengthened on the basis of the remarkable similarity of the rat CNR to CNVs recorded from primates (25). Hence, they may be used to model human neurobehavioral disorders in a way that is rarely possible. METHOD
Fourteen male, Sprague-Dawley rats were purchased as weanlings from commercial breeders (Harlan Sprague-Dawley). The animals were divided into two groups (7 animals per group). The control group was fed Purina Rat Chow, which contained a normal concentration of protein (23.4% protein; 4.5% fat; 25.3% crude fiber). The treatment group was fed a high-protein (HP) diet (50% protein; 5% fat; 5% crude fiber; 17.2% corn starch; 17.2% sucrose), which was commercially prepared by Research Diets, Incorporated. Both animal groups remained on the diets ad lib for 36-40 weeks, after which they were anesthetized and prepared for experimentation.
Preparation for Surgery On the day of experimentation, the animal was anesthetized using alpha-chloralose and urethane (50 mg/kg and 1.5 mg/kg IP, respectively). After induction of anesthesia, the hair over the animal's skull and along the back over the spine was clipped. The body temperature was maintained at 3 7 - 0 . 5 ° C by a heating pad placed under the animal and monitored by rectal probe. A surgical incision was made ventrally in the neck to expose the trachea. An incision was made in the trachea and PE 240 polyethylene tubing inserted into the airway to facilitate ventilation of the animal. The animal was permitted to breathe spontaneously.
Recording Evoked Potentials Evoked and event-related potentials were recorded using a Nicolet Compact Four Electrodiagnostic System. All records were averaged on-line. Each record was visually inspected for artifact before being stored on disk for off-line analysis.
Brainstem Auditory-Evoked Potentials (BAEPs) After the animal was anesthetized and surgically prepared as described above, its head was fixed in a stereotaxic frame (Stoelting). Ear bar adaptors were positioned in the indention of the animal's squamosal bones. Earplug speakers were inserted
into the left and right auditory meatus and secured m place with surgical tape. Platinum wire electrodes (Grass Instruments) ~tcre inserted subcutaneously over the skull. A reference electrude was positioned over the top of the skull at the midline /Cz!. ~\ctivc electrodes were positioned just behind each ear and lateral m the temporalis muscles (A1 and A2). A ground electrode was positioned over the frontal bone. The auditory stimulus consisted of rarifaction clicks, 100 microseconds in duration, and at a rate ot I 1.4/second. The stimulus was delivered monaurally at different intensity levels (30, 45, 60, 75, and 90 dB nHL). The opposite ear received noisemasking at 15 dB less than stimulus intensity. Two thousand sweeps were averaged, with a sweeptime of 5 milliseconds. The bandpass setting was 150-3 kHz. Two-channel recordings were obtained: Cz-AI, and Cz-A2.
Somatosensor 3' Evoked-Potentials (SS EPs ) A platinum wire electrode was inserted subcutaneously over the frontal bone (Fpz) for reference. Three platinum wire electrodes were used for recording: 1) over the midline of the parietal bone (Cz), 2) subdermally over the seventh lumbar spinous process (spL7), and 3) inserted into the base of the tail (T). A ground electrode was positioned subcutaneously over the interparietal bone. Stimulating electrodes consisting of two stainless steel pins were inserted into the tail of the rat (3 cm apart) and a ground electrode (steel pin) was inserted proximally, near the base of the tail. The stimulus consisted of an electrical shock delivered to the tail with a 200 microsecond duration and at a rate of 4.8/second. Stimulus intensity was adjusted to the minimum amperage required to observe a tail twitch response. Threechannel recordings were obtained; channel 1: Fpz-Cz, channel 2: Fpz-spL7, and channel 3: Fpz-T. For channel 1, the bandpass was 30-1.5 kHz. For channels 2 and 3, the bandpass was 5-1.5 kHz. A minimum of 2000 sweeps were averaged per record. with a sweeptime of 25 milliseconds.
Cortical Negativio' Responses (CNRs) The cranium was surgically exposed. Platinum needle electrodes were placed in contact with the dura mater through small holes made in the skull. A reference electrode was located over the frontal bone (2 mm anterior and 2 mm lateral to bregma). Two active electrodes (A3 and A4) were located over the parietal cortex (7.5 mm posterior, + 4 and - 4 mm lateral to bregma). A ground electrode was placed over the interparietal bone. Two-channel recordings were made (Fpz-A3, Fpz-A4l with a bandpass of 0.01-30 Hz. The alerting stimulus (AS) consisted of a continuous 1.5 kHz tone. The stimulus was delivered at a rate of 3/minute, with a 1 millisecond ramp, a 100 millisecond plateau, and at an intensity of 90 dB nHL. The rats were allowed to habituate, or become familiar with the alerting stimulus, in at least two sessions of exposure, with a minimum of 5 minutes between sessions. Each animal met criteria for habituation when the waveform recording demonstrated no positive deflection in response to AS (a startle response to the alerting stimulus). The recording electrodes were strategically positioned proximal to the auditory cortex (37) in order to facilitate recording of the startle response. This permitted a reliable determination of habituation without compromising the recording of expectancy negativity, which has both a frontal and parietal distribution (4,13). Eighty trials were averaged during each session, with a sweeptime of 5 seconds. All sweeps were externally triggered 1 second prior to AS. After the habituation sessions, AS was paired with the imperative stimulus (IS). To accomplish this, stimulating electrodes consisting of two stainless steel pins were inserted into the tail of the rat (3 cm apart). A ground electrode (steel pin) was inserted more proximally, near the base of the tail. IS consisted of an 11-volt, square wave pulse of 1.4
CORTICAL RESPONSES IN HIGH-PROTEIN DIET
889
second duration (Grass Stimulator, Model $48; Grass Stimulus Isolation Unit, Model SIU5B). The animal was trained with AS-IS in 3 sessions, with interstimulus intervals (ISI) of 0.5, 1.0, and 2.0 seconds, respectively. The intensity of the tail shock was kept constant for each session. Again, 80 trials were averaged in each session and intertrial intervals were at least 5 minutes apart. Following training to the AS-IS paradigm with 2-second ISI, another session was obtained in which IS was withheld, allowing the animal to undergo extinction of training. In a final session, the animal was retrained to AS-IS with a 2-second ISI.
I 1pV HABITUATION
+
EXTINCTION
Data Analysis RETRAINING
The latency of each component was measured from the onset of the stimulus to each deflection's highest point. Amplitude was calculated as the voltage difference measured peak-to-trough. The latencies and amplitudes of the peaks of the two diet groups were analyzed using two-tailed, unpaired Student's t-test. In all analyses, statistical significance was assumed when p < 0 . 0 5 was obtained.
AS I 0
I 1
t I z
I 3
I 4
I 5
SECONDS
RESULTS
Body weights were not different between the two groups at the time of sacrifice (Controls: 502---14 grams; HP group: 495--- 16 grams). On the day of experimentation, animals consuming the HP diet were observed to be more alarmed and resistive to handling than the counterparts on a normal diet, although these observations were not quantitated. Both groups, however, achieved an adequate level of anesthesia with the same dosages of urethane and alpha-chloralose. Figure 1 shows the results recorded from one animal during the CNR protocol. In the HP group, all animals habituated to the alerting stimulus within two sessions. Within the control group, 4 of the 7 animals required a third habituation session in order to meet criteria for habituation. When the imperative stimulus was paired with the alerting stimulus, all animals generated two distinct negative deflections during the interstimulus interval. Figure 2 shows the grand average CNR for one animal in the study. The first negative deflection (N1) was characterized by a sharp rise, early peak, and short duration. The second negative peak (N2) was characterized by a later peak, a slower rise time, and a much longer duration. N2 was occasionally bifurcated. The typical response to the imperative stimulus was a gradual, positive baseline shift in the waveform. After training with AS-IS with a 2-second interstimulus interval, the animals were allowed to undergo extinction reactions by discontinuing the imperative stimulus. Then each animal was retrained with AS coupled to IS, recreating the CNR. Absolute latencies, voltages, and durations for N1 and N2 are presented in Table 1. All data for Table 1 were calculated from the grand average waveforms, such as the one depicted in Fig. 2. N1 latencies, amplitudes, and duration were not statistically different between the two diet groups. There was a tendency for N1 amplitudes to be larger in the HP diet group, but the differences were not significant at the 95% confidence level. Regarding the N2 deflection, animals consuming the HP diet demonstrated significantly longer peak latencies (p<0.05). The HP group also demonstrated significantly larger N2 amplitudes (p<0.05). N2 durations were not different between the two groups. BAEPs were successfully recorded in all animals. Four positive deflections were reproducible. These were identified as Waves I through IV, typically seen in rodent short-latency auditory-evoked responses. Table 2 shows the absolute peak laten-
IS
t
FIG. 1. Cortical negativity responses (CNR) recorded from one urethane/ chloralose anesthetized rat in this study. The animal was conditioned using an alerting stimulus-imperative stimulus paradigm. Each EEG tracing is the average of 80 sweeps. In the top tracing, the animal was habituated to the alerting stimulus (AS), which was a tone cue. When the tone cue was followed by the imperative stimulus (IS), which was an electrical tail shock, there was a negative shift in the EEG baseline during the interstimulus interval. Withholding the IS resulted in extinction of the negative shift response. In the bottom tracing, the A 1 was coupled again to the IS, recreating the CNR.
cies of Wave IV recorded from the two diet groups, as well as the I-IV interpeak intervals. As can be seen, there were no differences between the two diet groups at any intensity level tested. Figure 3 shows the audiometric analyses of the animals in both diet groups. No reproducible peaks occurred when the animals were stimulated at 30 dB. Wave IV latencies were max-
I lpV +
N1
AS
IS
t I 0
I 1
NZ
t I Z
I 3
I 4
I 5
$ECONOS FIG. 2. Grand average (160 sweeps) of the two CNR responses depicted in Fig. 1. This technique improves the signal-to-noise ratio and more clearly defines two negative deflections (N1 and N2) generated by the frontal cortex when the animal has been conditioned to the alerting stimulus (AS) followed by an imperative stimulus (IS).
;',;90
B ROCK A N l) P R -\ S~,:\i
TABLE I PEAK LATENC1ES AND AMPLITUDES OF CNR PEAKS C o n t r o l Diet
N1 Latency(ms) (mean -+ SEM) Amplitude (IJ,V) (mean ± SEM) Duration (ms) (mean m SEM)
931 ± 93
High-Protein Diet
N2
NI
N2
1330 _+_ 1
1120 _+ 52
1560 ± 1"
0.75 ± 0.11
1.11 ± 0.13
1.74 ± 0.57
2.28 ± 0.25*
277 +_ 81
684 ± 91
244 ± 55
962 _+ 196
*Indicates p<0.05, control diet (N = 7) versus high-protein diet (N = 7), unpaired Student's t, two-tailed test. The data were measured from CNR grand averages for each animal, as depicted in Fig. 2.
imal at intensity levels at or above 75 dB. Also, the amplitudes of the different peaks were not different between the two groups. At 45 dB, Wave IV amplitudes averaged 0.04-+0.01 ixV for both diet groups. At 60 dB, Wave IV amplitudes in the control group were 0 . 1 3 ± 0 . 0 4 txV, versus 0 . 0 9 ± 0 . 0 2 txV in the HP group. At 75 dB, Wave IV amplitudes plateaued at 0.21 ± 0 . 0 6 I~V in the control group and 0 . 1 8 ± 0 . 0 3 txV in the HP diet group. Also, at 90 dB, I/IV amplitude ratios were not different between the two groups (control diet group: 4.47 -+ 1.02; HP diet group: 6.76-+1.69). In total, there were no significant differences in the BEAPs between the two diet groups. Somatosensory-evoked responses were also successfully recorded in all animals. Stimulus intensities required to observe a tail twitch response in the control diet group were 4.2 ± 0.9 mA, while the HP diet group required tail stimulus intensities of 3.7---0.6 mA. Tail stimulation resulted in 3 reproducible deflections in the waveforms, designated according to their peak latencies as N1.8, N5, and P16. In all animals, a negative deflection was most predominant in the Fpz-T channel, representing depolarization of the peripheral nerves in the tail. This deflection occurred with a peak latency of 1.86-+0.14 ms in the control animals, and 1.63 -+ 0.14 ms in the HP diet animals (not signifi-
4.500 O m
4.000 lip •
3.500
~
3.000 2.500
30
!
i
i
45
60
75
90
Click I n t e n s i t y (dB nHL) FIG. 3. Audiometric analyses of brainstem auditory evoked potentials recorded from animals fed the medium-protein diet (C, N = 7) and highprotein diet (HP, N = 7) (mean-+-SE). Wave IV response latency is plotted as a function of click intensity. Threshold for Wave IV generation was between 30-45 dB nHL. The slope of the latency-intensity function describes normal sensoryneural recruitment and information conduction.
cant). A second negative deflection was most easily identified in the Fpz-spL7 channel, and represented depolarization of the spinal cord, as the peripheral nerve enters the spinal cord and synapses. In the control animals, this negativity occurred with a peak latency of 5.44__.0.24 ms, and in the HP diet animals at 5.33-+0.10 ms (not significant). Peripheral conduction times were determined for the two groups by calculating the N1.8-N5 interpeak intervals, and were found to be not significantly different (control diet group: 3.58-+0.21 ms; HP diet group: 3.70-+0.15 ms). A positive deflection (P16) was identifiable in all 3 channels (all referenced to Fpz) which represented the response of the sensory cortex to the peripheral nerve stimulation. The absolute latencies and amplitudes of the PI6 responses recorded from the two diet groups are shown in Table 3, as well as calculated central conduction times (N5-P16 interpeak intervals). In total, there were no significant differences in the SSEPs between the two diet groups. DISCUSSION
The major observation made from this study was that animals fed a chronic, high-protein diet demonstrated CNRs which where significantly larger in amplitude and longer in latency than CNRs recorded from control animals (Table 1). The difference in CNR amplitude stands in contrast to the other recorded variables, BAEPs and SSEPs (Tables 2 and 3, respectively), which were not different between the two groups. It is important to note that differences between the diet groups are not explainable based upon differing levels of anesthesia. Although the animals consuming the HP diet acted more alarmed and resistive to handling than the control animals, both achieved an adequate level of anesthesia with the same dosages of urethane and alpha-chloralose. With the dosages used in this study, alpha-chloralose behaves as a sedative-hypnotic agent, while the urethane induces a profound and very long-lasting condition of anesthesia in the animal. The plane of anesthesia in each animal was monitored by toe clamp and whisker reflexes throughout the experiment. Urethane (ethyl carbamate) has often been the anesthetic of choice for neuroscientists, because it has less influence than other anesthetics on the normal activity of a wide variety of brain areas, including the cerebral cortex ( I l L In evoked potential studies, urethane anesthesia typically reduces amplitudes and increases latencies, but does not produce fundamental changes in wave morphology (2). Neither does the drug disrupt cortical mechanisms involved in learning, even at anesthetic dosages (25,26). Although the mechanism of action of urethane is unknown (11), the remarkable features of this drug
CORTICAL RESPONSES IN HIGH-PROTEIN DIET
891
TABLE 2 LATENCIES,AMPLITUDES,AND INTERPEAKINTERVALSOF BEAP PEAKS* Control Diet
Latency (ms) (mean - SEM) Amplitude (I~V) (mean --- SEM)
High-Protein Diet
Wave IV
I-IV Interval
Wave IV
I-IV Interval
2.94 _ 0.15
2.55 --- 0.13
2.86 --- 0.06
2.43 --- 0.05
0.24 --- 0.06
0.19 --- 0.04
*Recorded at 90 dB nHL. All data: p>0.05, unpaired Student's t, two-tailed test.
may be exploited to perform classical conditioning in a fully anesthetized animal. Wave I of the brainstem auditory potentials arises from depolarization of the auditory nerve; wave IV arises from the vicinity of the ipsilateral superior olive, and possibly as proximal as the lateral lemniscus; the I-IV interpeak interval is a measure of brainstem transmission time for the auditory information (31). Normal BEAPs indicate that the high-protein diet did not alter auditory information processing, at least at the level of the bralnstem. Similarly, normal SSEPs indicate that the high-protein diet had no obvious effect on the peripheral response to electrical stimulation of the tall. The P16 component is similar to the positive cortical potentials observed by others (14). This component is apparently generated by postsynaptic activation of cell bodies located in the cortex in response to thalamic input (1). The data suggests that the abnormal CNRs induced by the HP diet are not explained by changes in peripheral sensory mechanisms, such as alterations in sensory thresholds or defects in auditory acuity (24). The N1 component which preceded the CNRs in the two diet groups was not different (Table 1). This component is believed to be generated by the frontal cortex as an orienting response to the alerting stimulus (10) and reflects the processing of directional information (4). In primates, the N1 is not necessarily related to the contingency of the CNV paradigm (10). However, in the present study, generation of the frontal component was contingent with the imperative stimulus (Fig. 1). The rat frontal cortex responded to the AS-IS paradigm by generating a waveform that resembled primate frontal cortical potentials, but also demonstrated a contingency which resembled premotor cortical potentials. This observation is consistent with evidence that the
frontal cortex of the rat is not homologous to that of the primate, but is more similar anatomically to the primate cingulate and premotor cortices (17). Normal N1 responses reinforce the conclusion that auditory information processing in the HP diet group was not different from the control group. As mentioned previously, the CNR in the rat is analogous to the CNV recorded in primates, which strengthens our ability to understand the implications of these data. There is no doubt that the CNV amplitude depends upon the learning of the relevance of the AS-IS contingency (28). The precentral motor area is responsible for generation of this waveform and its appearance represents the preparation or execution of motor action by the subject (29). One important aspect of motor preparation is the condition of "motor set," which refers to the state of readiness of the cells of the motor cortex to generate a command for movement (38). In theory, cortical cell activity is determined at any moment by a previously learned motor program. The motor program may be changed on the basis of new sensory information and result in a different level of cell activity (38). Within the framework of this theory, the associative properties of AS processing changed the motor set of the rat cortical cells. Under the condition of the HP diet, amplification of the CNR may.be explained by an abnormality in the cellular process of motor programming that biased a larger area of cells toward synchronous depolarization. The data may also be explained by prolonged activation of the neural circuits involved in analyzing the associative properties of the alerting stimulus. The increase in CNR amplitude in the HP group was accompanied by an increase in absolute peak latency (Table 1). An increase in peak latency was unexpected, since rats on the HP diet demonstrated decreased response latencies to tall flick stim-
TABLE 3 LATENCIES,AMPLITUDES,AND INTERPEAKINTERVALSOF SSEP PEAKS* Control Diet N5-P16 P 16 Interval Latency (ms) (mean +__ SEM) Amplitude (I~V) (mean --- SEM)
16.03 ± 0.82 0.10 --- 0.02
10.58 -+ 0.82
High-Protein Diet N5-P16 P16 Interval 15.98 ± 0.56
10.68 ± 0.61
0.08 --- 0.01
*Recorded at 4.17 --- 0.95 mA (Control diet, N=7), 3.67 _+ 0.67 mA (HP diet, N=7). All data: p>0.05, unpaired Student's t, two-tailed test.
,~2
ulation (22). It was expected that motor response latency would be directly related to CNV peak latency. One possible explanation for this discrepancy is that CNV peak latency does not correlate to motor responses of the purely reflexive type. such as those elicited by the tail flick method. This observation underscores the fact that the CNV is uniquely related to motoric activity that involves the frontal cortex. Consider also that the CNV is a composite of negative deflections related to anticipation, cognition, and motor command, which may be evoked independently or in concert, according to experimental design (13). Measuring the latency of a single deflection in the CNV when a bifid wave is present may not be informative unless we are certain which feature of AS processing it represents. These data are in support of the theory that a chronic, highprotein diet results in an increase in preparatory arousal mechanisms. Tecce conceptualized "'arousal" as a hypothetical process which energizes behavior unselectively, affecting only intensity of response and being devoid of steering properties (28). Although there are cells in the premotor cortex which show directional specificity (38), the amplitudes of premotor potentials appear to correlate more strongly with the force of movement rather than the direction (4). This study sheds more light upon our earlier observations that rats fed a chronic, high-protein diet demonstrate behavioral hyperactivity and hyperresponsiveness to noxious stimuli (22). The present study would suggest that an alteration in central mechanisms is responsible for protein-induced hyperactivity and hyperresponsiveness. Also, the highamplitude CNRs observed in the HP diet group most likely represent an overexpression of motor cortical activity which is generated endogenously, not reflexively, since abnormalities in peripheral sensory mechanisms made no significant contribution to the observed behavior. These data also provide additional support for the suggestion that it is the protein content of the diet which correlated with the observed behavior, and not the carbohydrate content. Since carbohydrate diets increase tryptophan levels in the brain (33), which is known to decrease locomotor activity in rats (33,36), we would predict the carbohydrate effect on CNR amplitudes to be just the opposite of what was seen in this study. Therefore, it is unlikely that the carbohydrate content of the HP diet used in this study contributed to the high amplitude CNRs. Although dopamine was not measured in this study, the data are consistent with the theory that a high-protein diet stimulates dopaminergic activity, as suggested by the high amplitude CNRs (35). At this point, it is interesting that 4 of the 7 control animals required a third habituation session, whereas the HP animals habituated within two sessions. Although the experiments were not designed to precisely quantitate the onset of habituation, it is possible that the HP animals habituated more quickly than the control animals. Habituation normally involves the process of selective attention (7), and both are disrupted by dopaminergic stimulation (7,20). If the HP diet was stimulating a dopaminergic mechanism, then it should have taken longer for the HP animals to habituate. In the anesthetized animal which may not demonstrate selective attention, habituation may result from lower brain functions involved in attentional maintenance (9). Maintenance of attention is increased by L-DOPA (8) and is more directly correlated with basal arousal levels (28). The HP diet may have resulted in an enhancement of attentional maintenance (via increased basal arousal level) in such a way that preserved the anesthetized animal's ability to perform feature analysis of repetitive stimuli. This may have compensated for a delay in habituation otherwise seen in the anesthetized control animals. In humans, overexpression of dopaminergic mechanisms is
BROCK AN[) PRASAi).
involved in schizophrenia, a condition that is well characterized by low amplitude and prolonged CNVs ~35). In the presem study, animals fed a high-protein diet indeed had prolonge~ CNRs, but generated robust, rather than low-amplitude responses. Schizophrenics have normal BAEPs (24) but prolonged cortical SSEPs (3), both of which are normal in the high dietary protein group. Furthermore, schizophrenics more often present with either elevated or normal serum magnesium levels (5}. The electrophysiological effects of a high-casein diet presented herein do not fit very well with a model of schizophrenia. Also, the CNRs recorded from the HP diet group do not compare well to CNVs recorded from hyperactive humans. Such people present with a short latency, but rapidly deteriorating, frontal component in their CNVs, concomitant with an attention deficit, perhaps due to an impaired communication between the frontal cortex and thalamus (35). The frontal component in both diet groups demonstrated high variability, but the HP diet group NI showed a greater tendency to be prolonged compared to controls. High-amplitude CNVs similar to the CNRs observed in the high-protein group are seen in neurotic patients and in cases of psychosomatic illnesses, such as asthma (35). This raises the possibility that a dietary factor may be involved in these illnesses. A very important aspect of the rat CNR which must be mentioned is the fact that there is a cholinergic component, as well as a dopaminergic one, involved in the magnitude of the CNR (12,25). The frontal cortex in the rat receives its cholinergic innervation from the magnocellular forebrain nuclei (27). Presently, it is not clear what the relative roles are for the two neurotransmitter systems in the generation of eventcelated potentials. Others have shown that dopamine D 2 receptor activation has a facilitory effect on cholinergic depolarization of frontal cortical cells (39). The implication is that a high-protein diet causes an increase in dopaminergic activity which may have facilitated cholinergic activity in the frontal cortex, thus amplifying the negativity response. The cholinergic component of the CNR is modulated by the activity of gamma-aminobutyric acid (GABA), as shown by the fact that the negativity response in rats is partially inhibited by the microinjection of GABA into the magnocellular forebrain area (27). Investigators currently believe that diminished GABAergic activity in the cerebrum is an important mechanism involved in anxiety (21). Therefore, CNR recording in rats may have valuable applications in the study of panic disorder or other anxiety-related illnesses (18). Cortical negativity responses are only a component in the expression of motoric behavior, in which other brain regions are involved. In the present study, as in most CNV measurements, generation of the negativity response is constrained by the AS-IS paradigm. This cortical activity may be related to voluntary acts of the Process 1 type (19), which includes CNVs that are paced, or generated in association with sensory cues. With the present limits of knowledge, these must be held distinct from cortical activity of the Process II type, i.e., CNVs that are uniquely associated with fully independent, volitional acts (19). ACKNOWLEDGEMENTS This work was supported by the US Army Research and Development Command, grant DAMD-17-88-Z-8023. Opinions, interpretations, conclusions, and recommendations are those of the author and are not necessarily endorsed by the US Army. In conducting research using animals, the investigators adhered to the Guide for the Care and Use of Laboratory Animals, prepared by the Committee on Care and Use of Laboratory Animals of the Institute of Laboratory Animal Resources, National Research Council (NIH), publication #86-23, revised 1985.
CORTICAL RESPONSES IN HIGH-PROTEIN DIET
893
REFERENCES 1. Addae, J. I.; Stone, T. W. Purine receptors and kynnrenic acid modulate the somatosensory evoked potential in rat cerebral cortex. Electroencephalogr. Clin. Neurophysiol. 69:186-189; 1988. 2. Angel, A.; Gratton, D. A. The effect of anaesthetic agents on cerebral cortical responses in the rat. Br. J. Pharmacol. 76:541-549; 1982. 3. Boning, J.; Drechsler, F.; Neuhauser, B. Somatosensory event-related potentials and selective attention impairment in young chronic schizophrenics. Neuropsychobiology 21:146--151; 1989. 4. Bonnet, M.; Macday, W. A. Changes in contingent-negative variation and reaction time related to precueing of direction and force of a forearm movement. Brain Behav. Evol. 33:147-152; 1989. 5. Brackenridge, C. J.; Jones, I. H. Relation of hypermagnesemia to activity and neuroleptic drug therapy in schizophrenic states. J. Neurol. Neurosurg. Psychiatry 34:195-199; 1971. 6. Chafetz, M. D. Catecholamines and behavioral arousal. In: Chafetz, M. D., ed. Nutrition and neurotransmitters, the nutrient bases of behavior. Englewood Cliffs, NJ: Prentice-Hall, Inc.; 1990:67-87. 7. Cheal, M. L. Disruption of selective attention by apomorphine, but not amphetamine, in the Mongolian gerbil. Psychopharmacology (Berlin) 69:93-100; 1980. 8. Cheal, M. L. L-DOPA and piribedil alter different components of attentional behavior dependent on dose. Psychopharmacology (Berlin) 80:345-350; 1983. 9. Cheal, M. Attention and habituation: Separation of specific components in development. Physiol. Behav. 41:415-425; 1987. 10. Desmedt, J. E.; Debecker, J. Waveform and neural mechanism of the decision P350 elicited without pre-stimulus CNV or readiness potential in random sequences of near-threshold auditory clicks and finger stimuli. Electroencephalogr. Clin. Nenrophysiol. 47:648-670; 1979. ll. Dyer, R. S.; Rigdon, G. C. Urethane affects the rat visual system subanesthetic doses. Physiol. Behav. 41:327-330; 1987. 12. Ebenezer, I. Cortical slow potentials recorded in rats anaesthetized with urethane are affected by nicotine. Br. J. Pharmacol. 79:270P; 1983. 13. Frost, B. G.; Neill, R. A.; Fenelon, B. The determinants of the non-motoric CNV in a complex variable foreperiod, information processing paradigm. Biol. Psychol. 27:1-21; 1988. 14. Gilmore, R.; Brock, J.; Hermansen, M. C.; Baumann, R. Development of lumbar spinal cord and cortical evoked potentials after tibial nerve stimulation in the pre-term newborns: Effects of gestational age and other factors. Electroencephalogr. Clin. Neurophysiol. 68: 28-39; 1987. 15. Holl, M. G.; Allen, L. H. Comparative effects of meals high in protein, sucrose, or starch on human mineral metabolism and insulin secretion. Am. J. Clin. Nutr. 48:1219-1225; 1988. 16. Kanarek, R. B. Determinants of dietary self-selection in experimental animals. Am. J. Clin. Nutr. 42:940-950; 1985. 17. Kolb, B. Functions of the frontal cortex of the rat: A comparative review. Brain Res. Rev. 8:65-98; 1984. 18. Lader, M.; Mathews, A. Physiological changes during spontaneous panic attacks. J. Psychosom. Res. 14(4):377-382; 1970. 19. Libet, B.; Wright, E. W.; Gleason, C. A. Readiness-potentials preceding unrestricted "spontaneous" vs. pre-planned voluntary acts. Electroencephalogr. Clin. Neurophysiol. 54:322-335; 1982. 20. Matthysse, S. Dopamine and selective attention. In: Costa, E.; Gessa, G. L., eds. Advances in biochemical pharmacology, vol. 16. New York: Raven Press; 1977:667--669.
21. Nestoros, J. N. Gabaergic mechanisms and anxiety: An overview and a new neurophysiological model. Can. J. Psychol. 29:520-529; 1984. 22. Onaivi, E. S.; Brock, J. W.; Prasad, C. Behavior modification in rats involving changes in dietary protein. Nutr. Res., submitted; 1991. 23. Peters, J. C.; Harper, A. E. Adaptation of rats to diets containing different levels of protein: Effects on food intake, plasma and brain amino acid concentrations and brain neurotransmitter metabolism. J. Nutr. 115:382-398; 1985. 24. Pfefferbaum, A.; Horvath, T. B.; Roth, W. T.; Tinklenberg, J. R.; Kopell, B. S. Auditory brain stem and cortical evoked potentials in schizophrenia. Biol. Psychol. 15(2):209-223; 1980. 25. Pirch, J. H.; Corbus, M. J.; Rigdon, G. R. Event-related slow potentials in urethane anesthetized rats. Soc. Neurosci. Abstr. 9:1197; 1983. 26. Pirch, J. H.; Corbus, M. J.; Ebenezer, I. Conditioned cortical slow potential responses in urethane anesthetized rats. Int. J. Neurosci. 25:207-218; 1985. 27. Ridgon, G. C.; Pirch, J. H. Microinjection of procaine or GABA into the nucleus basalis magnocellularis affects cue-elicited unit responses in the rat frontal cortex. Exp. Neurol. 85:283-296; 1984. 28. Rockstroh, B.; Elbert, T.; Birbaumer, N.; Lutzenberger, W. Slow brain potentials and behavior. Baltimore, MD: Urban and Schwarzenberg; 1982. 29. Rohrbaugh, J. W.; Syndulko, K.; Sanquist, T. F.; Lindsley, D. B. Synthesis of the contingent negative variation brain potential from noncontingent stimulus and motor elements. Science 208:11651168; 1980. 30. Rosenkilde, C. E. Functions of the prefrontal cortex: Behavioral investigations using ablation and electrophysiological techniques in rats, cats, dogs and monkeys. Acta Physiol. Scand. (Suppl. 514):658; 1983. 31. Shaw, N. A. The auditory evoked potential in the rat--a review. Prog. Neurobiol. 31:19--45; 1988. 32. Sos, J.; Szelenyi, I. Diets for animal experiments. Budapest: Academiai Kiado; 1974. 33. Spring, B. Effects of foods and nutrients on the behavior of normal individuals. In: Wurtman, R. J.; Wurtman, J. J., eds. Nutrition and the brain, vol. 7. New York: Raven Press; 1986:1-47. 34. Stallings, V. A.; Archibald, E. H.; Pencharz, P. B. Potassium, magnesium, and calcium balance in obese adolescents on a proteinsparing modified fast. Am. J. Clin. Nutr. 47:220-224; 1988. 35. Timsit-Berthier, M. Contingent negative variation (CNV) in psychiatry. In: MeCallum, W. C.; Zappoli, R.; Denoth, F., eds. Cerebral psychophysiology: Studies in event-related potentials. EEG (Suppl. 3):429-434; 1986. 36. Warbritton, J. D.; Stewart, R. M.; Baldessarini, R. J. Decreased locomotor activity and attenuation of amphetamine hyperactivity with intraventricular infusion of serotonin in the rat. Brain Res. 143:373-382; 1978. 37. Webster, W. R. Auditory system. In: Paxinos, G., ed. The rat nervous system, vol. 2. Sydney, Australia: Academic Press; 1985:153184. 38. Wise, S. P. The primate premotor cortex: Past, present, and preparatory. Annu. Rev. Neurosci. 8:1-19; 1985. 39. Yang, C. R.; Mogenson, G. J. Dopaminergic modulation of cholinergic responses in rat medial prefrontal cortex: An electrophysiological study. Brain Res. 542:271-281; 1990.