Pii: S0304-3940(00)01437-3

Preceding or immediately following fear-conditioning rats were exposed for 30 min to either a sham ®eld, one of two symmetrical (sine-wave 7 , 20 Hz) magnetic ®elds or to one of two complex magnetic ®elds whose waveforms were modeled after salient electrophysiological patterns within either the hippocampal formation (theta-burst) or the amygdaloid complex (burst-®ring). The magnetic ®elds were presented in successive 2 s intervals through each of the three spatial planes and then simultaneously within all three planes. Field strengths ranged between 0.5 and 1 microTesla. Only the group exposed after the conditioning to the theta-burst (hippocampal) magnetic ®elds displayed evidence of forgetting, as inferred by their marked attenuation of freezing behavior, during contextual extinction 24 h later. This powerful treatment explained 75% of the variance in the extinction scores. Behavioral responses to the discrete conditioned stimulus were not affected. These ®ndings are consistent with the involvement of the hippocampus in learned fear to contextual stimuli but not to discrete auditory stimuli and suggest that physiologically relevant stimuli may be delivered to the brain by weak, complex magnetic ®elds whose intensities are ubiquitous within modern environments. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Magnetic ®elds; Conditioning; Fear responses; Complex wave forms; Theta bursts We have pursued the hypothesis that the brain functions as an electromagnetic neuromatrix. From this perspective: (1) all behaviours are the products of neuroelectromagnetic patterns associated with neurochemical activity and (2) brain space is immersed in complex time-varying electromagnetic ®elds that can be diffuse or constrained within non-adjacent spatial increments. For an environmental stimulus from natural (e.g. geomagnetic) or synthetic (electronic) sources to intercalate with the neuromatrix, the temporal and spatial characteristics of processes inside and outside of the brain must be congruent. The results of the present experiments showed that an electromagnetic pulse, whose intrinsic shape and interstimulus interval had been speci®cally designed to in ̄uence an electrophysiological process involved with learning, can evoke qualitatively conspicuous changes in behavior. Maintained immobility by Rattus norvegicus while in the presence of contextual (non-speci®c or background) cues or of discrete (tone) cues previously associated with an aversive event has been employed as an inference of learning [1]. Extensive research indicates that contextual and discrete classical conditioning involve different pathways within the brain [4]. Lesions of the hippocampus selectively impair acquisition and consolidation of contextual fear conditioning but minimally in ̄uence the acquisition and consolidation of auditory-cued fear behavior [10]. On the other hand, functional lesions of the amygdala impair acquisition and consolidation of all fear-related behavior [10]. Consequently fear conditioning to contextual or to discrete stimuli provides a behavioral paradigm by which contributions from the hippocampus and amygdala may be functionally dissociated. During learning many amygdaloid neurons exhibit burst ®ring [15] whereas hippocampal pyramidal neurons display three to ®ve pulse-bursts within the range of theta rhythms [8]. Synaptic facilitation or long term potentiation (LTP), a major electrophysiological correlate of long term memory, can be easily induced in vivo when `natural patterns' are applied through direct electrical stimulation [2]. We Neuroscience Letters 292 (2000) 99±102 0304-3940/00/$ see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S0304-3940(00)01437-3 www.elsevier.com/locate/neulet * Corresponding author. Fax: 11-705-671-3844. E-mail address: [email protected] (M.A. Persinger). reasoned that pre-conditioning and post-conditioning application of weak complex magnetic ®elds, similar to prominent hippocampal electrophysiological patterns, should selectively impair the acquisition and the consolidation of contextual fear conditioning. Ninety-nine male Wistar strain albino rats served as subjects. Animals were obtained from Charles River (Quebec) and were 8±12 months old. Rats were housed in groups of 2±3 per cage in standard colony housing. Ambient temperature was maintained at 20 ^ 18C. The light±dark cycle was 12:12 with photophase onset at 07:30 h local time. Food and water were available ad libitum. All conditioning and testing took place during the mid to late photophase. Contextual and discrete conditioning as well as contextual extinction testing were conducted in a 28 £ 20 £ 20 cm modi®ed Plexiglas and aluminum operant chamber. The raised ̄oor of the chamber consisted of 18 steel rods spaced 1.5 cm apart. Each rod was connected to a unit designed to deliver scrambled electric shocks. A 6 cm diameter speaker was mounted in one lateral wall and was connected to a frequency generator programmed to emit a speci®c tone. To facilitate removal of animals, a non-conducting frame whose dimensions approximated those of the chamber was inserted into the chamber prior to conditioning and testing each animal. Ambient sound intensity was approximately 50 dB. Extinction testing of the discrete stimulus was assessed in another room in a novel apparatus (a 33 £ 21 £ 23 Plexiglas chamber with acoustic equipment mounted in one end). All apparatus was cleaned and dried before each animal was trained or tested. Magnetic ®elds were applied in a room that had not been employed for conditioning. The computer equipment utilized to generate the magnetic ®elds converted a series of values ranging from 0 through 255 to a voltage ranging from ±5 to 15 V. Neutral polarity was represented by the value 128 while negative and positive polarities were represented by values less than 128 and greater than 128, respectively. This series of values, if plotted against their sequence in the computer ®le, graphically depicted the waveform of the magnetic ®eld. A Zenith XT personal computer with a port latency of 140ms propagated the values in the waveform ®le to a custom constructed, digital-to-analogue converter that transformed the ®le into the appropriate voltage. The apparatus that emitted the magnetic ®elds as a function of the voltage-graded input was a 30 £ 30 £ 30 cm Plexiglas box with 50 ohm solenoids af®xed perpendicularly to the center of each external face of the cube (six solenoids total). The solenoids were connected to the digital-to-analogue converter. The mechanical component from each solenoid had been replaced with a 15 £ 0:5 cm steel rod to focus the weak-intensity magnetic ®elds. For this application the digital-to-analogue converter included a solid-state switching device (switching time in the order of nanoseconds) which enabled successive activation of opposing solenoid pairs (one north pole, one south pole) for 0.5 s each and culminated with the activation all three solenoid pairs for 0.5 s (see Fig. 1). As a result, the ®eld was rotated through three-dimensional Cartesian space every 2 s (e.g. 0.5 Hz) and then through all three planes simultaneously. These rotations, completed every 8 s, occurred continuously throughout the 30 min treatment period. The bottom onethird of the apparatus had been ®lled with corncob bedding to ensure the animal's resting position was in the center of the applied magnetic ®eld. Five treatment conditions were employed for the present study. Subjects were exposed to sham conditions, to one of the two complex magnetic ®eld patterns, or to one of two sinusoidal magnetic ®eld patterns. The ®rst complex pattern was a theta burst stimulation pattern (maximum intensity about 500 nT) designed to mimic the ®ring parameters of hippocampal pyramidal cells during learning and to induce long-term potentiation or LTP in hippocampal slices [12]. This pattern consisted of ®ve-pulse bursts at 100 Hz separated by 140 ms (e.g. theta rhythm). The second complex magnetic ®eld pattern was a burst-®ring pattern [11] modeled after recordings of amygdaloid activity in epileptic patients; its average intensity was approximately 900 nT. The two sinusoidal patterns had frequencies (maximum intensities in parentheses) of 7 Hz (400 nT) and 20 Hz (900 nT) and were employed as intensity controls for the theta burst stimulation and burst-®ring pulses. Rats that received sham conditions (,10 nT) were placed in the treatment apparatus with only the digital-to-analogue converter activated. Analysis of variance (ANOVA) was employed as the primary statistical tool. Post-hoc analyses were completed with the Scheffe (P , 0:01) test. All analyses were completed using SPSS software on a VAX 4000 computer. Eta-squared values were employed as indicators of effect size. It is de®ned as the amount of variance explained in the dependent measure, the immobility scores, by manipulations of the independent variable, magnetic ®eld patterns. B.E. McKay et al. / Neuroscience Letters 292 (2000) 99±102 100 Fig. 1. Drawing of the exposure apparatus showing the positions of the pairs of modi®ed solenoids along each spatial dimension. See body of text for details. In Experiment I, 31 rats were trained in the contextual conditioning paradigm. Each animal was placed in the conditioning chamber and received four unsignaled footshocks (1-s, 0.5 mA) at 60 s intervals subsequent to an initial 2 min habituation period. The rat was removed from the conditioning apparatus immediately following the last footshock presentation and placed in the treatment room. Animals were randomly assigned to each of the ®ve treatment groups (sham, hippocampal pattern, amygdaloid pattern, 7 Hz sine wave, or, 20 Hz sine wave). Subsequent to 30 min of magnetic ®eld (or sham) exposure the rats were returned to their home cages. Freezing to contextual cues was assessed approximately 24 h later when the animals were returned to the conditioning apparatus and tested singly. The rats' behaviours were scored nominally for the presence or absence of freezing via a time-sampling procedure for 1 s every 8 s throughout the 8 min extinction test [6]. The number of freezing observations was converted to a percentage of total observations. One-way analysis of variance indicated a statistically signi®cant difference between treatment conditions for the proportion of freezing behaviours during contextual extinction testing (F…4; 26† ˆ 19:45, P , 0:001, eta-squaredˆ 75%). Post hoc analysis revealed the source of the effect to be the signi®cantly reduced freezing-to-context for rats exposed to the theta burst stimulation magnetic ®eld relative to all other treatment conditions that did not differ signi®cantly from each other. The results are shown in Table 1. Pre-existing group differences in motoric activity, as de®ned by the numbers of forward movements of the head and forepaws over the midline of the chamber, were not evident (F…4; 26† ˆ 1:07, P ˆ 0:39). In Experiment II, 30 rats were exposed for 30 min to either sham, burst-®ring or theta-burst patterned magnetic ®elds prior to contextual conditioning. Because the two sinusoidal magnetic ®eld treatments did not affect contextual extinction testing in Experiment I they were excluded from additional experiments. All conditioning parameters were identical to Experiment I. One-way analysis of variance demonstrated no statistically signi®cant group differences (F…2; 27† ˆ 2:28, P ˆ 0:12). The means and standard errors of the mean (SEM) are also shown in Table 1. In Experiment III, 21 rats were conditioned to a discrete auditory stimulus. Following a 2 min pre-shock period these animals received four pairings of a 10 s auditory stimulus (2000 Hz, 90 dB) that terminated with the onset of a 1-s 0.5 mA footshock. The inter-pairing interval was 60 s. Rats were immediately relocated to the treatment room and exposed for 30 min to either the sham-®eld or to one of the two complex magnetic ®elds. Rats were tested for their recall of the discrete stimulus in the novel environment approximately 24 h later. Rats were placed in this environment for a total of 10 min which consisted of a 2 min pretone period and an 8 min tone presentation period. Freezing was scored during the tone presentation using the same procedure outlined for Experiment I. A one-way analysis of variance showed there were no statistically signi®cant differences between the treatment groups for the duration of freezing behavior during extinction to the discrete stimulus (F…2; 18† ˆ 1:12, P ˆ 0:35). Experiment IV was procedurally identical to Experiment III except magnetic ®elds were applied for 30 min prior to discrete conditioning and 17 rats were employed as subjects. There were no statistically signi®cant differences between the three groups in freezing behavior during extinction (F…2; 14† ˆ 0:50, P ˆ 0:62). Differences in pre-tone freezing immediately prior to tone extinction as a function of treatment group were not signi®cant statistically (F…2; 15† ˆ 0:02, P ˆ 0:98). Differences in spontaneous locomotor activity in all rats (n ˆ 47) receiving pre-conditioning magnetic ®eld treatment were not evident (F…2; 44† ˆ 1:80, P . 0:05). The results of these four experiments have shown that weak complex magnetic ®elds, patterned after the temporal parameters that elicited LTP in hippocampal slices [12], interfered with the memory of the contextual cues but not the discrete cues after fear conditioning when the rats were tested 24 h later. The effect was strongest when the rats were exposed to this speci®c ®eld, but not to a less symmetric burst-®ring pattern, during the 30 min after the acquisition. Exposure to the hippocampal pattern prior to conditioning attenuated freezing, i.e. exhibited a mild interference with contextual memory. However the effect was ambiguous statistically. Sinusoidal magnetic ®elds, a popular temporal structure in many studies involving bioelectromagnetic effects, had no statistically signi®cant in ̄uence upon the behavioral response to the conditioned contextual cues. Temporally symmetrical (sine-wave or square wave) magnetic ®elds may require higher intensities to evoke signi®cant behavioral changes because they may be more dependent upon induction of electrical currents within susceptible structures or processes. That the sine wave magnetic ®elds were presented at the same intensity as the complex patterned ®elds yet did not affect this type of memory suggests waveB.E. McKay et al. / Neuroscience Letters 292 (2000) 99±102 101 Table 1 Means and SEM for percent freezing to contextual stimuli in rats exposed to sham conditions, one of two complex magnetic ®elds, or one of two sinusoidal magnetic ®elds for 30 min immediately following or preceding the conditioning session Treatment Post-Conditioning Pre-Conditioning