- Email: [email protected]
The relationship between isometric force requirement and forelimb tremor in the rat
Physiology & Behavior 69 (2000) 285–293
The relationship between isometric force requirement and forelimb tremor in the rat John A. Stanforda, Elena Vorontsovad, Stephen C. Fowlerb,c,d,* a
Department of Anatomy and Neurobiology, University of Kentucky, HDFL/4001 Dole, Lexington, KY 40536, USA Departments of bHuman Development and Family Life and cPharmacology and Toxicology, d Life Span Institute, University of Kansas, Lawrence, KS 66045, USA Received 22 June 1999; received in revised form 11 October 1999; accepted 9 December 1999
Abstract To explore the effects of isometric force of rodent forelimb contraction on forelimb tremor, rats were trained to press downward on an isometric force transducer to raise a water-filled dipper cup and maintain force to keep the dipper in the raised position while licking. Force requirements were then manipulated parametrically to measure the effects of escalating force output on forelimb tremor and other variables. In the Peak-Force greater than Hold-Force (PF ⬎ HF) manipulation, the forces required to raise the dipper were 20, 40, and 60 g (each condition for about 2 weeks), while the force required to maintain the dipper in the raised position remained 6.7 g for all three conditions. In the Peak-Force equal to the Hold-Force (PF ⫽ HF) manipulation, rats were required to maintain the “dipper-raising” force throughout the response. The forces required were 20 g, 40 g, and 60 g (each for 2 weeks). For all force requirement manipulations, data were analyzed within and across conditions. As expected, force output increased with increased force requirements. Spectral analysis of force–time records revealed that during all manipulations, high-frequency (⬎10 Hz) forelimb tremor increased with increased force output, an effect that is consistent with human studies, and that may reflect increases in the number of motor units firing at higher rates. Additionally, with the exception of the 60-g PF ⫽ HF condition, there were within-condition decreases in tremor and increases in task engagement, evidence suggesting increased muscle strength as a function of experience (i.e., “physical training”). Taken together, the results suggest that the rodent-based method may provide a valuable, noninvasive functional assay for animal models of disorders that affect skeletal muscle control in humans. © 2000 Elsevier Science Inc. All rights reserved. Keywords: Isometric; Force; Forelimb; Tremor; Operant
1. Introduction There are two physiological mechanisms through which the mammalian nervous system increases skeletal muscle force output to meet increasing force demands: by increasing (a) the number of motor units participating in a muscle contraction (i.e., recruitment), or (b) the firing rate of individual motor units participating in a contraction (i.e., firing rate modulation) [1,2]. Previous studies with healthy human subjects have shown that increases in peripheral muscle tremor accompany increased muscle force output [3–6]. The portion of force-augmented tremor under 30 Hz has been attributed to muscle force variations resulting from the contractile twitches of individual motor units firing asynchronously between the modes of recruitment (ca. 6–8 Hz) and complete fusion (between 25 and 30 Hz) [3]. This normal relationship between muscle force and the behavior of mo* Corresponding author. Tel.: 785-864-0715; Fax: 785-864-5202 E-mail address: [email protected]
tor units has been demonstrated to be altered in patients suffering from various diseases that affect motor systems [4]. For example, in Parkinson’s disease, cerebellar diseases, and chronic degenerative demyelinating neuropathies, the main power in muscle force power spectra is shifted to frequency ranges that are lower than those in normals, while the main power is shifted to higher frequency ranges in individuals suffering from axonal neuropathies, motor neuronopathies, myopathic disorders, and lesions of the upper motor neurons. These alterations represent impairments in neural force-producing mechanisms, and the measurement of them is useful for diagnostic and pathophysiological purposes [4]. For this reason, corresponding measurement of these phenomena in nonhumans would be a valuable advance in experimental and preclinical studies of diseases that affect motor control in humans. The relationship between force output and tremor has been assessed using a variety of techniques (e.g., surface EMGs, accelerometers, acoustic myography). It has been suggested that the spectral analysis of force variation during
0031-9384/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S0031-9384(99)00 2 4 8 - 6
286
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
isometric contractions is an ideal method for the concurrent measurement of force-related tremor and motor unit ensemble activity [4]. Indeed, a substantial consistency between spectral analysis of tremor, EMG activity, and muscle vibration signals has been demonstrated [5]. Because of the general agreement between the results obtained through these methods, the measurement of force-related tremor via isometric force transducers affords an opportunity to conduct cross-species studies of this phenomenon more efficiently and under relatively noninvasive conditions (i.e., without muscle implants or attached accelerometers or electrodes). To our knowledge, the relationship between voluntary force output and tremor has heretofore not been quantified in rats. The primary barrier to such studies is perhaps the tractability of the subjects. Therefore, evaluation of voluntary force-related tremor in nonhumans (as well as in nonverbal humans) requires the use of operant behavioral techniques [7]. Operant studies have been undertaken with rats in which force output and tremor were utilized as dependent variables. For example, we have previously demonstrated forelimb hypertonia, force discontrol, and increases in rats’ forelimb tremor produced by known tremorogenic agents (e.g., neuroleptics and cholinergic agonists), as well as forelimb hypotonia and decreases in tremor produced by drugs known to decrease tremor in the clinic (e.g., some atypical neuroleptics and anticholinergic drugs) [8–10]. Conversely, several studies have been conducted to assess the effect of forelimb force as an independent variable in the context of drug- and nondrug-related operant behavior [11,12]. Parametric studies of the relationship between voluntary isometric force output and tremor in rats, however, have been lacking. In the present study, this relationship was assessed in a rat-based forelimb tremor task [9]. Rats were trained to press downward on an isometric force transducer with one forelimb for water reinforcement such that as long as a criterion force was maintained, a dipper cup remained raised so that the rat could lick the water. Force–time records were Fourier transformed and spectrally analyzed to assess the effects not only of increasing force requirements on forelimb tremor, but also the effect of training on this relationship. Two hypotheses were put forth and tested. The first was that the relationship between force and tremor is conserved across mammalian species, and that tremor would increase concurrently with increased force output as it does in humans. The second was that the relationship between isometric force output and tremor would be influenced by training under the task’s demands.
2. Materials and methods 2.1. Subjects Twenty-four male Sprague–Dawley rats (Harlan, Indianapolis, IN) with a mean body weight of 349.5 g (SEM 5.9 g) on Day 1 of the baseline condition served as subjects. Rats were restricted to 10 min of water each day following
experimental sessions, and were given Purina Rat Chow ad lib via food hoppers attached to their cages. This regimen permitted a slow weight gain of 4 to 5 g/week. The vivarium was maintained on a 12-h light–dark cycle (lights on from 0800–2000 h), with temperature maintained at 26⬚C. Experimental sessions were conducted between approximately 1300 and 1600 h. Use of the animals was approved by the University of Kansas IACUC, and procedures adhered to the NIH Guide for the Care and Use of Laboratory Animals. 2.2. Apparatus Training and experimental sessions were conducted in four operant chambers that were enclosed in individual sound-attenuated chambers as described elsewhere [8]. In each chamber a cylindrical recession provided access to the 0.5-mL cup of a solenoid-operated dipper. In the front panel of each chamber was a rectangular aperture located such that a rat could reach with its forelimb the operandum located outside the chamber. The operandum was an 18-mm diameter disk rigidly attached to Model 31 load cells (0–250 g range, Sensotec, Columbus, OH), which continuously measured the force exerted on the disk by the forelimb. The flat surface of the disk was oriented at 0⬚ from the vertical axis (i.e., parallel to the floor of the chamber). The spatial arrangement of the aperture, the manipulandum, and the raised dipper cup was such that the trained rat could press the manipulandum with a single extended forepaw and drink from the dipper cup at the same time (for specific dimensions of apparatus, see [9]). The force transducer–operandum complex and the signal-conditioning electronics were interfaced to a 386-based IBM compatible computer that sampled transducer output at 100 samples/s and provided a force resolution of 0.33 g equivalent weights. In the unloaded state the transducer– operandum system had a natural frequency (resonant frequency) of about 160 Hz; resonance artifact of the unloaded system was eliminated by a 100 Hz low-pass filter. Oscilloscope measurement of the natural frequency of the force transducer–operandum system with a 50-g mass resting on the operandum yielded an estimate of 30 Hz. According to physical principles, the frequency of a spring mass system is inversely related to the mass of the system: higher mass reduces frequency of oscillation. Thus, with spring stiffness held constant, the natural frequency of the force transducer– operandum system declines as the mass of the operandum increases. When the rat’s forelimb actively engages the operandum, this subject-generated force could conceivably affect the natural frequency of the recording system, even though very substantial dampening is afforded by the glabrous tissue of the forepaw. Because the natural frequency with a 50-g load was 30 Hz, it is unlikely that any of the frequency shifts seen in the power spectra below this frequency were related to limb loading effects; i.e., limb-loading effects were generally less than 50 g. For these reasons only frequencies of 25 Hz and below were analyzed. Force
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
was expressed in gram-equivalent weight units instead of Newtons because gram weights were used in the daily static calibration procedure. 2.3. Procedure 2.3.1. Baseline Following water restriction, rats were manually trained to press down on the operandum to obtain water reinforcement. One-half of the rats were required to use their right forelimb while the other half were required to use their left forelimb. The reinforcement contingency was controlled by computer software such that when the rat produced a minimum initial response force of 20 g (i.e., the peak-force requirement), the water cup was raised and presented through the hole in the cylindrical recession. The cup remained raised as long as the rat maintained a force of at least 6.7 g (i.e., the hold-force requirement). This hysteresis was introduced so that the rat would release its hold upon the operandum to refill the dipper cup, thereby producing relatively discrete responses. The hysteresis also provided an opportunity for the rat to rest briefly between responses. The response force peaked and decreased early in the first second after response initiation, whereupon it quickly decreased and plateaued for the duration of the response, remaining at a relatively constant force value well above the 6.7-g dipper-sustaining criterion. It generally took about 6 s from response initiation for the rat to completely lick the water from the cup, the rat’s posture remaining fixed throughout its response. Daily sessions of responding were 8 min 11 s. 2.3.2. Manipulations for peak force greater than hold force (PF ⬎ HF) Three peak-force requirements were used: 20.0, 40.0, and 60.0 g, each for 13 daily sessions. The dipper-sustaining force was 6.7 g in each case. Hereafter, the force criteria will be designated by the dipper-raising force followed by the dipper-sustaining force, for example, 20.0/6.7. Following the 60/6.7 phase, the rats were returned to the 20.0/6.7 requirement for 4 days. 2.3.3. Manipulations for peak force equal to hold force (PF ⫽ HF) After the PF ⬎ HF phase, rats were tested under force criteria that required them to maintain the dipper-raising force throughout the response (i.e., the force criterion for dipper maintenance was the same as for dipper activation). The three successive force requirement conditions were 20.0/20.0, 40.0/40.0, and 60.0/60.0 g, lasting 14, 15, and 14 days, respectively. Following the 60.0/60.0 phase, a second return-to-baseline phase ensued, during which the original 20.0/6.7 force criteria were used for 6 days. 2.3.4. Daily alternation of force requirement (DAFR) Following the PF ⫽ HF phase, effects of daily alternation of 20.0/6.7 and 40.0/13.0 force criteria were evaluated. This was done for two reasons. The first reason was to determine whether the rats could make rapid adjustments in
287
force output in response to changed force requirements, presumably by correlating kinesthetic feedback with exteroceptive feedback. The second reason was to include a condition in which amount of practice was limited to 1 day instead of 14 days for a given force requirement (i.e., to try to separate amount of tremor from physical training effects). For this condition, rats were divided into two groups of 12 rats each. Group 1 started this phase under the 20.0/6.7 criteria, while group 2 started under the 40.0/13.0 criteria. On day 2, the criteria reversed for the two groups (such that group 1 performed under the 40.0/13.0 criteria and group 2 performed under the 20.0/6.7 criteria). This cycle repeated for 13 days until it was determined by the experimenters that the differences between the two force conditions were not great enough to produce robust separation between the two force criteria (i.e., the rats were not discriminating the differences). To make the daily difference in force criteria more discriminable, the higher requirement was set at 60.0/ 20.0. Sixteen days of responding ensued under these conditions (i.e., 20.0/6.7 and 60.0/20.0 on alternate days). 2.4. Dependent variables Experimental sessions for each rat were analyzed by computer programs that provided several measures of behavior. Off-line, raw force–time data were parsed into individual waveforms (identified by computer as response bouts consisting of at least 4.36 s of dipper-up responding). Individual waveforms of responses (i.e., a response was one drinking bout of about 6 s) were ensemble averaged in the time domain for the first 4.36 s of the event. Two measures were derived from these averaged waveforms: the maximum force attained by the rat during the first second (max force), and the mean force sustained during the remaining 3.36 s (hold force). To quantify isometric force-related forelimb tremor, the 3.36-s hold segment of each individual force–time waveform was subjected to a prime factor fast Fourier transformation (Alligator Technologies, Fountain Valley, CA) to yield separate power spectra for each event. The spectra were then ensemble averaged in the frequency domain for each subject. Integrated power in the 10- to 25-Hz frequency band was taken as the sum (i.e., integral) of the frequency estimates lying within the 10–25 Hz frequency band, and was the area under the curve (power within a given bandwidth) when examining power spectra. Lower frequencies were not included in the tremor measurement because previous analyses showed that frequencies in the region of 7 Hz are the results of the rat’s lapping behavior (rhythmic tongue movements) during the task, and the aim of the research was to quantify forelimb tremor, not movements of the tongue. In addition to the force and tremor measures, time on task (TOT), which is analogous to operant response rate and is the cumulative time of operandum contact during the 8-min session measured to the nearest second, was also cal-
288
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
culated for each rat. TOT was defined in terms of a 1-g forelimb force threshold (i.e., any touching of the operandum counted even if it did not produce water).
Table 1 Means (bold type) and standard errors of the mean (below means) for the indicated dependent variables for the first and last days of exposure to the three different force requirements during the PF ⬎ HF manipulations
2.5. Quantitative analysis For statistical analyses, data for each dependent variable were taken from the initial and final days of each manipulation. Data were analyzed using a two-way, repeated-measures factorial analysis of variance (ANOVA) using SYSTAT’s Multivariate General Linear Model. One factor was the “practice-effects” factor (first versus last day of each phase), and the other was the factor for the required force condition. Data for the PF ⬎ HF manipulations, the PF ⫽ HF manipulations, and the DAFR manipulation were analyzed separately. To quantify the effect of practice on the relationship between isometric force output and tremor across the force requirement manipulations, linear regression was used; log power was regressed on mean hold force separately for the first and last days of performance for each of the parameter values of the PF ⬎ HF and PF ⫽ HF conditions. 3. Results 3.1. PF ⬎ HF conditions Means and standard errors of means for dependent variables measured during this condition are presented in Table 1. Figure 1 illustrates the effects of increasing the peakforce requirement on the shape of one rat’s raw force–time waveforms, and the top panel of Fig. 2 (which includes averaged force–time waveforms for all three PF ⬎ HF conditions) illustrates these effects on the averaged waveforms for all rats (see Figs. 1 and 2, top panel). As expected, max force increased as the peak-force criteria were increased, F(2, 44) ⫽ 144.775, p ⬍ 0.001; this measure did not significantly change from the first day to the final day of the phases. Although rats were not required to increase their dipper-maintaining hold forces during this phase, the hold portion of their responses also increased with increases in the dipper-raising force requirement (see Fig. 2, top panel). This increase in hold force as a function of dipper-raising force criterion was also significant, F(2, 44) ⫽ 42.239, p ⬍ 0.001. Like max force, hold force did not significantly change between the first and last days (see Table 1). Power spectral analyses revealed that integrated power across almost the entire 0–25-Hz frequency range increased with increased force output (see Fig. 2, bottom panel). This was confirmed for the integrated power in the 10–25-Hz band, as the measure increased significantly with the increased force requirements, F(2, 44) ⫽ 31.061, p ⬍ 0.001 (see Table 1). There was a significant practice effect for integrated power in the 10–25-Hz band, as (with the exception of the 60.0/6.7 condition) the measure decreased from the first day to the final day of each 2-week practice period, F(1, 22) ⫽ 6.432, p ⬍ 0.05. With respect to task engage-
20.0/6.7
40.0/6.7
Variable
First
TOT (s)
285.32 289.56 279.20 306.60 300.23 307.35 8.18 4.97 13.15 8.33 10.08 9.53 31.96 32.58 40.92 41.54 54.60 55.58 1.41 1.59 1.07 1.14 1.25 1.14 24.73 25.79 28.38 28.99 31.53 33.24 1.26 1.33 1.60 1.49 1.51 1.52
Max Forcea,b (g) Hold Forceb (g) Integrated power in the 10–25 Hz bandb,c (log10 power)
4.66 0.14
Last
4.46 0.16
First
60.0/6.7
4.87 0.15
Last
4.72 0.13
First
5.09 0.13
Last
5.08 0.11
a
The values for max force are taken from the averaged force-time waveforms. Due to variation in latencies to reach peak forces in individual responses, the values for this dependent variable were lower than the actual peak forces emitted. b Significant force-requirement effect, F-test, df ⫽ 2,44, p ⬍ 0.01. c Significant practice effect, F-test, df ⫽ 1,22, p ⬍ 0.05.
ment as measured by TOT, neither force requirement nor practice significantly influenced the measure during any of the PF ⬎ HF manipulations. 3.2. PF ⫽ HF conditions Means and standard errors of means for dependent variables measured during this condition are presented in Table 2. Due to nonresponding by six rats on the final day of the 60.0/60.0 phase, 18 rats were included in the ANOVA equations during the PF ⫽ HF conditions for all dependent variables except for the time-on-task measures (where nonresponders produced a value of 0). Figure 1 illustrates the way in which increasing the PF ⫽ HF requirements influenced the shape of one rat’s raw force–time waveforms. The averaged force–time waveforms depicted in the top panel of Fig. 3 shows the effect of the force requirements on the force– time waveforms of the entire group. Increases in the peak force and hold force requirements produced significant increases in both max force, F(2, 34) ⫽ 174.412, p ⬍ 0.001 and hold force, F(2, 34) ⫽ 153.066, p ⬍ 0.001. In addition, there was a significant practice effect for hold force, as the measure generally increased from the beginning to the end of each practice period, F(1, 17) ⫽ 23.906, p ⬍ 0.001. A significant practice-by-force requirement interaction affirmed that the increases in hold force as a function of practice were greater as the force requirements were increased, F(2, 34) ⫽ 5.382, p ⫽ 0.009. Max force did not significantly change with practice (i.e., first-day vesus last-day comparison). Integrated power in the 10–25-Hz band also increased as a function of increases in peak force and hold force requirements, F(2, 34) ⫽ 77.910, p ⬍ 0.001. This variable did not significantly change within the PF ⫽ HF 2-week periods. Generally, the PF ⫽ HF manipulations produced an in-
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
289
crease in TOT from the 20.0/20.0 phase to the 40.0/40.0 phase, then a decrease when the contingencies were increased to 60.0/60.0, which is supported by a significant practice-by-force-requirement interaction, F(2, 46) ⫽ 3.236, p ⬍ 0.05. 3.3. DAFR Condition Means and standard errors of means for dependent variables measured during this condition are presented in Table 3. Comparison of the averaged force–time waveforms for the 20.0/6.7 condition in Fig. 4 with Fig. 2 reveals that rats’ force output during this later occurring DAFR condition had increased since the PF ⬎ HF conditions. Inspection of Table 3 suggests that the force increase was a result of training prior to this phase and not the alternating force requirements during this phase, because the increase was present at the beginning of the DAFR condition, and there was no practice effect during this condition. Notwithstanding the baseline increase in force output for the 20.0/6.7 conditions, ANOVAs revealed significant required force effects for max force, F(1, 22) ⫽ 36.999, p ⬍ 0.001; and hold force, F(1, 22) ⫽ 24.17, p ⬍ 0.001 during the daily alternation procedure (see Fig. 4, top panel). In addition to increases in max force and hold force, ANOVAs revealed a significant increase in integrated power in the 10–25-Hz band as a function of force requirement, F(1, 22) ⫽ 24.383, p ⬍ 0.001 (see Table 3). Although TOT did not vary significantly as a function of required force, the measure did increase throughout the manipulation, yielding a significant practice effect, F(1, 23) ⫽ 16.37, p ⫽ 0.001. 3.4. Effects of practice on force-related tremor The effects of practice on the relationship between isometric force output and tremor are illustrated in Fig. 5. The regression lines for the first and last days of practice during each force-requirement phase are highly log linear and approximately parallel, with the tremor power on the last day falling consistently below the tremor power for the first day of data for each force requirement, even though the hold force rises somewhat (e.g., compare the circle and square at the upper right of both functions) during the 2 weeks of
Fig. 1. Force–time waveforms from a representative rat during portions of the final sessions of each of three types of force requirement conditions: (A) peak force greater than hold force; (B) peak force equal to hold force; (C) daily alternation of force requirement. For each condition, time proceeds from the left to the right of each row and from the bottom row to the top row. Force is plotted on the y-axis, where the vertical line on the left end of the bottom row of each graph represents 60.0 g. For each condition, as the lines get darker, higher force requirements are denoted (see legends). This rat was already engaged in the task with the forelimb maintaining downward force on the operandum and drinking water when the plot began. Note the regular force oscillations during the relatively flat (i.e., zero slope ) portions of the records.
290
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293 Table 2 Means (bold type) and standard errors of the mean (below means) for the indicated dependent variables for the first and last days of exposure to the three force requirements during the PF ⫽ HF manipulations 20.0/20.0
40.0/40.0
60.0/60.0
Variable
First
First
First
TOTc (s)
256.21 265.09 254.65 285.99 253.66 227.60 11.65 10.30 14.48 15.49 22.88 27.59 42.33 42.18 59.39 60.12 72.45 77.46 1.86 1.91 1.25 2.26 1.65 2.27 33.55 34.97 49.27 52.48 61.81 70.85 1.39 1.40 1.58 2.15 3.06 2.74
Max Forcea,b (g) Hold Forceb,c,d (g) Integrated power in the 10–25 Hz bandb (log10 power)
5.14 0.10
Last
4.92 0.13
5.94 0.11
Last
5.73 0.12
6.48 0.15
Last
6.56 0.08
The values for max force are taken from the averaged force-time waveforms. Due to variation in latencies to reach peak forces in individual responses, the values for this dependent variable were lower than the actual peak forces emitted. a Significant force-requirement effect, F-test, df ⫽ 2,34, p ⬍ 0.01. b Significant practice effect, F-test, df ⫽ 1,17, p ⬍ 0.01. c Significant practice-by-force requirement interaction, F-test, df ⫽ 2,46, p ⬍ 0.05. d Significant practice-by-force requirement interaction, F-test, df ⫽ 2,34, p ⬍ 0.01.
4. Summary
Fig. 2. Averaged force–time waveforms (top panel) and power spectra (bottom panel) for all rats during the final sessions of the PF ⬎ HF manipulations. Note that the peaks for the averaged force–time waveforms for the 40.0/6.7 and 60.0/6.7 conditions are lower than the required peak force criteria. This is a result of averaging waveforms with varying latencies to reach the actual peak in each individual response.
To summarize, rats were able to emit the forelimb forces that were required during the PF ⬎ HF and PF ⫽ HF manipulations. Forelimb forces generally increased from the beginning to the end of each manipulation (i.e., with practice). Increased force requirements did not significantly decrease rats’ engagement in the task. Indeed, task engagement measures increased with practice during each phase. As isometric force output increased, forelimb tremor increased concurrently. However, with the exception of the highest force conditions (i.e., 60.0/6.7 and 60.0/60.0), forelimb tremor power decreased with practice. 5. Discussion
training. In each of the six cases (three PF ⬎ HF and three PF ⫽ HF conditions) the ratio of tremor power to mean hold force was lower on the last day compared to the first day (p ⬍ 0.05, Wilcoxon rank sum test). This result indicates that tremor power decreased with practice (or physical training) when the level of the hold force was held constant. Note the rightward shift of the squares versus the circles in Fig. 5. This shift establishes that, while increased tremor accompanied increased isometric force output, with practice, higher forces were attained without increases in tremor. Inspection of the first and last days’ values for the daily alternation data (triangles) suggests that physical training effects did not continue in this condition, probably because substantially higher force requirements had preceded this phase of the work.
We report here increases in integrated power in the 10– 25-Hz frequency band (i.e., isometric muscle tremor) as a result of increases in voluntary force output in rats. Previous studies with healthy human subjects have demonstrated that increases in isometric force output elevate power in the frequency range that was elevated in the present study [3–6]. In an earlier review, it was argued that the spectral analysis of force records provides valuable information about not only increases in tremor, but also about the pooled activity of motor units as a result of synchronization and recruitment [1]. Moreover, it has also been suggested that the spectral analysis of muscle force activity provides information regarding asynchronous firing of muscle units (and, therefore, recruitment) that is superior to information provided by surface EMG analysis, because the higher frequencies of the
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
291
Table 3 Means (bold type) and standard errors of the mean (below means) for the indicated dependent variables for the first and last days of exposure to the condition involving daily alternation of force requirements 20/6.7
60/20
Variable
First
Last
First
Last
TOTc (s)
303.94 7.119 47.208 2.66 39.195 2.527
331.759 7.673 47.401 2.757 39.828 2.734
290.543 12.557 64.396 1.895 50.69 2.18
333.463 10.885 62.472 1.836 51.73 2.395
5.097 0.163
5.095 0.182
5.775 0.103
5.742 0.106
Max forcea,b (g) Hold forceb (g) Integrated power in the 10–25 Hz Bandb (log10 power) a
The values for Max Force are taken from the averaged force-time waveforms. Due to variation in latencies to reach peak forces in individual responses, the values for this dependent variable were lower than the actual peak forces emitted. b Significant force requirement effect, F-test, df ⫽ 1,22, p ⬍ 0.01. c Significant practice effect, F-test, df ⫽ 1,23, p ⬍ 0.05.
Fig. 3. Averaged force–time waveforms (top panel) and averaged power spectra (bottom panel) for all rats during the final sessions of the PF ⫽ HF manipulations. See Fig. 2 caption for further explanation.
EMG spectrum may be confounded by covarying electrical phenomena such as the conduction velocities of muscle fibers [4]. Indeed, the EMG power spectrum is influenced by not only muscle fiber conduction velocities, but also by the motor unit action potential, the number of motor units firing in proximity to the electrode, and the recording conditions [13]. Therefore, because of the number of variables that affect EMG, and because of the reported correlation between the results produced through these two methods [5], for motor unit phenomena ranging between 1 and 25 Hz, the spectral analysis of isometric muscle force variation may be a desirable and efficient alternative to more invasive methods (e.g., the use of intramuscular or surface electrodes or accelerometers). We have determined the peak in the power spectrum around 7 Hz to be attributable to rats’ licking frequencies (unpublished observations; force exerted by the tongue on the water cup was measured simultaneously with the forelimb force). In addition to this spectral peak, there are two
other less prominent peaks in the spectra—one at around 13 Hz, and one at around 20 Hz (e.g., see bottom axes in Figs. 2, 3, and 4). In a study examining frequency peaks during human finger muscle contraction, one group reported prominent peaks at 10, 20, and 40 Hz, and attributed their regularity to rhythmic output from the central nervous system [5]. This group argued that, because of the regularity of these peaks under varying force conditions, a central oscillator may be functioning to provide “timing” signals in the form of frequency-coded muscle commands. This argument is consistent with current theories regarding the role of rhythmic activity in the cerebellum in coordinating cell ensembles as they govern skeletal muscles [14]. The fact that there was a peak at 20 Hz in the present study suggests that whatever underlies this peak in humans may also produce this peak in rats. Similarly, although the peak around 13 Hz does not exactly match the 10-Hz peak reported in humans, it could also be an example of a rhythm generator conserved across species, only in the present case it could have been influenced (i.e., shifted rightward) by the prominent peak around 7 Hz. This explanation raises the issue of whether the 13 and 20 Hz peaks observed in this study represent harmonics of the dominant 7-Hz peak. Although we cannot entirely rule out this possibility, previous data from our laboratory support the relative independence of these peaks. In a study in which low doses of the tremorogenic agent harmaline was administered to rats in this procedure, we reported substantial increases in the power of the frequency spectrum around 13 Hz in the absence of increases in the height of the peak around 7 Hz [15]. Interestingly, however, the peaks associated will the 7-, 13-, and 20-Hz regions of the power spectrum were all shifted leftward by harmaline. Therefore, determination of the degree to which the phenomena reported in this study are analogous to human studies awaits
292
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
Fig. 5. Group mean integrated power in the 10–25 Hz band (tremor) as a function of mean hold force during the PF ⬎ HF, PF ⫽ HF, and DAFR manipulations for the first and last days of experience with each of the different force requirements. Regression lines show the relationship between isometric force output and tremor early (the first day, upper line) and late (last day, lower line) in training with the PF ⬎ HF and PF ⫽ HF manipulations (data for the DAFR conditions were not included in the regression line calculations). Note that higher isometric forces were attained with less forelimb tremor after 2 weeks of training, except in the alternation procedure.
Fig. 4. Averaged force–time waveforms (top panel) and power spectra (bottom panel) for all rats during the final sessions of the procedure involving daily alternation of the force requirements (DAFR). See Fig. 2 caption for further explanation.
the analysis of force recordings in rats that do not contain the 7-Hz peak (i.e., in a pressing-without-licking preparation). The fact that force-related increases in the power spectrum were not limited to the frequencies between 10 and 25 Hz, especially in the PF ⫽ HF condition, raises the question of whether the increases in power observed here are merely attributable to the increased amplitude of the force signals (i.e., instead of changes in the behavior of motor units). Although we cannot rule this possibility out, our demonstration that, with practice, increased force output was accompanied by decreased tremor, suggests otherwise. Moreover, we have previously reported tremorogenic drug-induced increases in the heights of various peaks of the power spectra in the absence of increases in hold force [10,15]. Taken together, these findings support our hypothesis that the phenomena measured in this study were specific to isometric force-related changes in motor unit ensemble activity.
In the present study, there was an overall tendency for practice to affect the relationship between isometric forelimb force output and tremor. This effect was manifested by the generation of higher forces while tremor generally decreased. Although we expected the force–tremor relationship to be influenced by practice, previous studies examining the effects of isometric training on EMG measures have demonstrated training-induced increases in motor unit activity [16]. Two physiological adaptations have been proposed to contribute to training-related increases in force output: muscle hypertrophy and neural adaptation (in the form of the recruitment of a “functional reserve” of previously inactive motor units with training) [16]. Although alterations in motor unit activity likely account for the observed force requirement effects on tremor, our results suggest that the ability of our rats to emit higher forces with diminished tremor over time may have resulted from training-induced muscle hypertrophy. The results in the present study are based on forces generated by rats under voluntary, sustained force output conditions in which a downward press with a single designated forelimb was required. Rats are capable of producing forces in excess of those required in this study by emitting either ballistic operandum striking behaviors (e.g., [11,17]) or rapid grasp-and-pull behaviors (e.g., [18]). Although these behaviors can yield valuable information, there are several reasons why sustained, nonballistic responses such as those used in the present study are valuable. The first reason is
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
that, to spectrally analyze force recordings so that biophysical phenomena ranging from low frequency (e.g., licking) to high frequency (e.g., muscle unit activity) can be assessed, sustained force output (i.e., output enduring for several seconds) is required to provide adequately long records for meaningful Fourier analysis. In the present case, the power spectrum ranges from 0 to 25 Hz. This means that, according to a common rule of thumb stating that frequency resolution is generally the reciprocal of the observed time interval [19], time samples of at least 1 s must be recorded (and, indeed, samples of greater durations than the minimum based on this rule allow for increased accuracy of estimation and reliability). Samples longer than 1 s cannot be accomplished with a rat’s ballistic responses. Another rationale for utilizing sustained isometric forelimb force is that, although tremor and other neuromuscular phenomena can be measured by using nonisometric responses (e.g., [5]), the signals recorded from moving muscles contain frequencies that result from mechanical properties of these movements [4,5]. Regarding force differentiation as measured in the daily alteration procedure (DAFR), the fact that rats were generally able to produce appropriate force output as evidenced by the significant main effects for force requirement during this phase, suggests that rats were able to discriminate the force requirements and adjust their behavior accordingly when the differences between the requirements were suitably large (20.0/6.7 versus 60.0/20.0). In a study examining the effects of different response topographies on force discrimination in rats, we previously demonstrated that rats were better able to confine their responses to force bands (lower and upper force requirements) when the required response was a downward press than when the response was a grasp and pull in the horizontal plane [18]. We concluded that the grasp-and-pull topography may have elicited unconditioned species-typical feeding, burrowing, or scratching tendencies that interfered with the conditioning of circumscribed force output. Because the present contingencies did not include force “bands” (i.e., there was no upper force limit), the present findings suggest that rats tended to minimize response cost by somewhat limiting their force output during the low-force conditions. The results revealed interesting information regarding the effects of varying force requirements on fine motor behavior in rats. The measurement of voluntary force output, as well as high-frequency variations in muscle force (i.e., tremor) as a result of voluntary force output, from rats provides a homologue for similar studies being conducted with humans [4,5]. The use of the forelimb tremor task procedure with rats could provide valuable, noninvasive, functional assays for nonhuman models of peripheral muscular and neuromuscular disorders, such as the 6-OHDA–induced Parkinsonism model or the sciatic nerve constriction neuropathy model. Furthermore, creative programming and utilization of concurrent task engagement and force mea-
293
sures may also allow for cross-species usage of terms such as maximal voluntary force (and percentages thereof) in studies assessing isometric training in physical therapeutic contexts.
Acknowledgments This research was supported by MH43429.
References [1] Freund H-J. Motor unit and muscle activity in voluntary motor control. Physiol Rev 1983; 63:387–436. [2] Fuglevand AJ. Neural aspects of fatigue. Neuroscientist 1996;2:203–6. [3] Allum JHJ, Dietz V, Freund H-J. Neuronal mechanisms underlying physiological tremor. J Neurophysiol 1978;41:557–71. [4] Homberg V, Reiners K, Hefter H, Freund H-J. The muscle activity spectrum: spectral analysis of muscle force as an estimator of overall motor unit activity. Electroencephalogr Clin Neurophysiol 1986;63:209–22. [5] McAuley JH, Rothwell JC, Marsden CD. Frequency peaks of tremor, muscle vibration and electromyographic activity at 10 Hz, 20 Hz and 40 Hz during human finger muscle contraction may reflect rhythmicities of central neural firing. Exp Brain Res 1997;114:525–41. [6] Orizio C, Perini R, Diemont B, Figini MM, Veicsteinas A. Spectral analysis of muscular sound during isometric contraction of bicepts brachii. J Appl Physiol 1990;68:508–12. [7] Newland MC. Operant behavior and the measurement of motor dysfunction. In: Weiss B, O’Donoghue J, editors. Neurobehavioral Toxicity: Analysis and Interpretation. New York: Raven Press, 1995. pp. 273–97. [8] Fowler SC, Davison KH, Stanford JA. Unlike haloperidol, clozapine slows and dampens rats’ forelimb force oscillations and decreases force output in a press-while-licking behavioral task. Psychopharmacology (Berlin) 1994;116:19–25. [9] Fowler SC, Liao R-M, Skjoldager P. A new rodent model for neuroleptic-induced pseudo-parkinsonism: low doses of haloperidol increase forelimb tremor in the rat. Behav Neurosci 1990;104:449–56. [10] Stanford JA, Fowler SC. Scopolamine reversal of tremor produced by low doses of physostigmine in rats: evidence for a cholinergic mechanism. Neurosci Lett 1997;225:157–60. [11] Alling K, Poling A. The effects of differing response-force requirements on fixed-ratio responding of rats. J Exp Anal Behav 1995;63:331–46. [12] Makhay M, Alling K, Poling A. Effects of cocaine on fixed-ratio responding of rats: modulation by required response force. Pharmacol Biochem Behav 1994;48:511–4. [13] Fuglsang–Frederiksen A, Ronager J. The motor unit firing rate and the power spectrum of EMG in humans. Electroencephalogr Clin Neurophys 1988;70:68–72. [14] Welsh JP, Lang EJ, Sugihara I, Llinas R. Dynamic organization of motor control within the olivocerebellar system. Nature 1995;374:453–7. [15] Stanford JA, Fowler SC. At low doses, harmaline increases forelimb tremor in the rat. Neurosci Lett 1998;241:41–4. [16] Bandy WD, Hanten WP. Changes in torque and electromyographic activity of the quadriceps femoris muscles following isometric training. Phys Ther 1993;73:455–67. [17] Elsner J, Fellmann C, Zbinden G. Response force titration for the assessment of the neuromuscular toxicity of 2,5-Hexanedione in rats. Neurotoxicol Teratol 1988;10:3–13. [18] Kirkpatrick MA, Fowler SC. Operant force-band differentiation by rats using two different response topographies. Bull Psychonom Soc 1989;27:52–4. [19] Marple SL, Jr. Digital Spectral Analysis With Applications. Englewood Cliffs, NJ: PTR Prentice Hall, 1987.
The relationship between isometric force requirement and forelimb tremor in the rat John A. Stanforda, Elena Vorontsovad, Stephen C. Fowlerb,c,d,* a
Department of Anatomy and Neurobiology, University of Kentucky, HDFL/4001 Dole, Lexington, KY 40536, USA Departments of bHuman Development and Family Life and cPharmacology and Toxicology, d Life Span Institute, University of Kansas, Lawrence, KS 66045, USA Received 22 June 1999; received in revised form 11 October 1999; accepted 9 December 1999
Abstract To explore the effects of isometric force of rodent forelimb contraction on forelimb tremor, rats were trained to press downward on an isometric force transducer to raise a water-filled dipper cup and maintain force to keep the dipper in the raised position while licking. Force requirements were then manipulated parametrically to measure the effects of escalating force output on forelimb tremor and other variables. In the Peak-Force greater than Hold-Force (PF ⬎ HF) manipulation, the forces required to raise the dipper were 20, 40, and 60 g (each condition for about 2 weeks), while the force required to maintain the dipper in the raised position remained 6.7 g for all three conditions. In the Peak-Force equal to the Hold-Force (PF ⫽ HF) manipulation, rats were required to maintain the “dipper-raising” force throughout the response. The forces required were 20 g, 40 g, and 60 g (each for 2 weeks). For all force requirement manipulations, data were analyzed within and across conditions. As expected, force output increased with increased force requirements. Spectral analysis of force–time records revealed that during all manipulations, high-frequency (⬎10 Hz) forelimb tremor increased with increased force output, an effect that is consistent with human studies, and that may reflect increases in the number of motor units firing at higher rates. Additionally, with the exception of the 60-g PF ⫽ HF condition, there were within-condition decreases in tremor and increases in task engagement, evidence suggesting increased muscle strength as a function of experience (i.e., “physical training”). Taken together, the results suggest that the rodent-based method may provide a valuable, noninvasive functional assay for animal models of disorders that affect skeletal muscle control in humans. © 2000 Elsevier Science Inc. All rights reserved. Keywords: Isometric; Force; Forelimb; Tremor; Operant
1. Introduction There are two physiological mechanisms through which the mammalian nervous system increases skeletal muscle force output to meet increasing force demands: by increasing (a) the number of motor units participating in a muscle contraction (i.e., recruitment), or (b) the firing rate of individual motor units participating in a contraction (i.e., firing rate modulation) [1,2]. Previous studies with healthy human subjects have shown that increases in peripheral muscle tremor accompany increased muscle force output [3–6]. The portion of force-augmented tremor under 30 Hz has been attributed to muscle force variations resulting from the contractile twitches of individual motor units firing asynchronously between the modes of recruitment (ca. 6–8 Hz) and complete fusion (between 25 and 30 Hz) [3]. This normal relationship between muscle force and the behavior of mo* Corresponding author. Tel.: 785-864-0715; Fax: 785-864-5202 E-mail address: [email protected]
tor units has been demonstrated to be altered in patients suffering from various diseases that affect motor systems [4]. For example, in Parkinson’s disease, cerebellar diseases, and chronic degenerative demyelinating neuropathies, the main power in muscle force power spectra is shifted to frequency ranges that are lower than those in normals, while the main power is shifted to higher frequency ranges in individuals suffering from axonal neuropathies, motor neuronopathies, myopathic disorders, and lesions of the upper motor neurons. These alterations represent impairments in neural force-producing mechanisms, and the measurement of them is useful for diagnostic and pathophysiological purposes [4]. For this reason, corresponding measurement of these phenomena in nonhumans would be a valuable advance in experimental and preclinical studies of diseases that affect motor control in humans. The relationship between force output and tremor has been assessed using a variety of techniques (e.g., surface EMGs, accelerometers, acoustic myography). It has been suggested that the spectral analysis of force variation during
0031-9384/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S0031-9384(99)00 2 4 8 - 6
286
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
isometric contractions is an ideal method for the concurrent measurement of force-related tremor and motor unit ensemble activity [4]. Indeed, a substantial consistency between spectral analysis of tremor, EMG activity, and muscle vibration signals has been demonstrated [5]. Because of the general agreement between the results obtained through these methods, the measurement of force-related tremor via isometric force transducers affords an opportunity to conduct cross-species studies of this phenomenon more efficiently and under relatively noninvasive conditions (i.e., without muscle implants or attached accelerometers or electrodes). To our knowledge, the relationship between voluntary force output and tremor has heretofore not been quantified in rats. The primary barrier to such studies is perhaps the tractability of the subjects. Therefore, evaluation of voluntary force-related tremor in nonhumans (as well as in nonverbal humans) requires the use of operant behavioral techniques [7]. Operant studies have been undertaken with rats in which force output and tremor were utilized as dependent variables. For example, we have previously demonstrated forelimb hypertonia, force discontrol, and increases in rats’ forelimb tremor produced by known tremorogenic agents (e.g., neuroleptics and cholinergic agonists), as well as forelimb hypotonia and decreases in tremor produced by drugs known to decrease tremor in the clinic (e.g., some atypical neuroleptics and anticholinergic drugs) [8–10]. Conversely, several studies have been conducted to assess the effect of forelimb force as an independent variable in the context of drug- and nondrug-related operant behavior [11,12]. Parametric studies of the relationship between voluntary isometric force output and tremor in rats, however, have been lacking. In the present study, this relationship was assessed in a rat-based forelimb tremor task [9]. Rats were trained to press downward on an isometric force transducer with one forelimb for water reinforcement such that as long as a criterion force was maintained, a dipper cup remained raised so that the rat could lick the water. Force–time records were Fourier transformed and spectrally analyzed to assess the effects not only of increasing force requirements on forelimb tremor, but also the effect of training on this relationship. Two hypotheses were put forth and tested. The first was that the relationship between force and tremor is conserved across mammalian species, and that tremor would increase concurrently with increased force output as it does in humans. The second was that the relationship between isometric force output and tremor would be influenced by training under the task’s demands.
2. Materials and methods 2.1. Subjects Twenty-four male Sprague–Dawley rats (Harlan, Indianapolis, IN) with a mean body weight of 349.5 g (SEM 5.9 g) on Day 1 of the baseline condition served as subjects. Rats were restricted to 10 min of water each day following
experimental sessions, and were given Purina Rat Chow ad lib via food hoppers attached to their cages. This regimen permitted a slow weight gain of 4 to 5 g/week. The vivarium was maintained on a 12-h light–dark cycle (lights on from 0800–2000 h), with temperature maintained at 26⬚C. Experimental sessions were conducted between approximately 1300 and 1600 h. Use of the animals was approved by the University of Kansas IACUC, and procedures adhered to the NIH Guide for the Care and Use of Laboratory Animals. 2.2. Apparatus Training and experimental sessions were conducted in four operant chambers that were enclosed in individual sound-attenuated chambers as described elsewhere [8]. In each chamber a cylindrical recession provided access to the 0.5-mL cup of a solenoid-operated dipper. In the front panel of each chamber was a rectangular aperture located such that a rat could reach with its forelimb the operandum located outside the chamber. The operandum was an 18-mm diameter disk rigidly attached to Model 31 load cells (0–250 g range, Sensotec, Columbus, OH), which continuously measured the force exerted on the disk by the forelimb. The flat surface of the disk was oriented at 0⬚ from the vertical axis (i.e., parallel to the floor of the chamber). The spatial arrangement of the aperture, the manipulandum, and the raised dipper cup was such that the trained rat could press the manipulandum with a single extended forepaw and drink from the dipper cup at the same time (for specific dimensions of apparatus, see [9]). The force transducer–operandum complex and the signal-conditioning electronics were interfaced to a 386-based IBM compatible computer that sampled transducer output at 100 samples/s and provided a force resolution of 0.33 g equivalent weights. In the unloaded state the transducer– operandum system had a natural frequency (resonant frequency) of about 160 Hz; resonance artifact of the unloaded system was eliminated by a 100 Hz low-pass filter. Oscilloscope measurement of the natural frequency of the force transducer–operandum system with a 50-g mass resting on the operandum yielded an estimate of 30 Hz. According to physical principles, the frequency of a spring mass system is inversely related to the mass of the system: higher mass reduces frequency of oscillation. Thus, with spring stiffness held constant, the natural frequency of the force transducer– operandum system declines as the mass of the operandum increases. When the rat’s forelimb actively engages the operandum, this subject-generated force could conceivably affect the natural frequency of the recording system, even though very substantial dampening is afforded by the glabrous tissue of the forepaw. Because the natural frequency with a 50-g load was 30 Hz, it is unlikely that any of the frequency shifts seen in the power spectra below this frequency were related to limb loading effects; i.e., limb-loading effects were generally less than 50 g. For these reasons only frequencies of 25 Hz and below were analyzed. Force
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
was expressed in gram-equivalent weight units instead of Newtons because gram weights were used in the daily static calibration procedure. 2.3. Procedure 2.3.1. Baseline Following water restriction, rats were manually trained to press down on the operandum to obtain water reinforcement. One-half of the rats were required to use their right forelimb while the other half were required to use their left forelimb. The reinforcement contingency was controlled by computer software such that when the rat produced a minimum initial response force of 20 g (i.e., the peak-force requirement), the water cup was raised and presented through the hole in the cylindrical recession. The cup remained raised as long as the rat maintained a force of at least 6.7 g (i.e., the hold-force requirement). This hysteresis was introduced so that the rat would release its hold upon the operandum to refill the dipper cup, thereby producing relatively discrete responses. The hysteresis also provided an opportunity for the rat to rest briefly between responses. The response force peaked and decreased early in the first second after response initiation, whereupon it quickly decreased and plateaued for the duration of the response, remaining at a relatively constant force value well above the 6.7-g dipper-sustaining criterion. It generally took about 6 s from response initiation for the rat to completely lick the water from the cup, the rat’s posture remaining fixed throughout its response. Daily sessions of responding were 8 min 11 s. 2.3.2. Manipulations for peak force greater than hold force (PF ⬎ HF) Three peak-force requirements were used: 20.0, 40.0, and 60.0 g, each for 13 daily sessions. The dipper-sustaining force was 6.7 g in each case. Hereafter, the force criteria will be designated by the dipper-raising force followed by the dipper-sustaining force, for example, 20.0/6.7. Following the 60/6.7 phase, the rats were returned to the 20.0/6.7 requirement for 4 days. 2.3.3. Manipulations for peak force equal to hold force (PF ⫽ HF) After the PF ⬎ HF phase, rats were tested under force criteria that required them to maintain the dipper-raising force throughout the response (i.e., the force criterion for dipper maintenance was the same as for dipper activation). The three successive force requirement conditions were 20.0/20.0, 40.0/40.0, and 60.0/60.0 g, lasting 14, 15, and 14 days, respectively. Following the 60.0/60.0 phase, a second return-to-baseline phase ensued, during which the original 20.0/6.7 force criteria were used for 6 days. 2.3.4. Daily alternation of force requirement (DAFR) Following the PF ⫽ HF phase, effects of daily alternation of 20.0/6.7 and 40.0/13.0 force criteria were evaluated. This was done for two reasons. The first reason was to determine whether the rats could make rapid adjustments in
287
force output in response to changed force requirements, presumably by correlating kinesthetic feedback with exteroceptive feedback. The second reason was to include a condition in which amount of practice was limited to 1 day instead of 14 days for a given force requirement (i.e., to try to separate amount of tremor from physical training effects). For this condition, rats were divided into two groups of 12 rats each. Group 1 started this phase under the 20.0/6.7 criteria, while group 2 started under the 40.0/13.0 criteria. On day 2, the criteria reversed for the two groups (such that group 1 performed under the 40.0/13.0 criteria and group 2 performed under the 20.0/6.7 criteria). This cycle repeated for 13 days until it was determined by the experimenters that the differences between the two force conditions were not great enough to produce robust separation between the two force criteria (i.e., the rats were not discriminating the differences). To make the daily difference in force criteria more discriminable, the higher requirement was set at 60.0/ 20.0. Sixteen days of responding ensued under these conditions (i.e., 20.0/6.7 and 60.0/20.0 on alternate days). 2.4. Dependent variables Experimental sessions for each rat were analyzed by computer programs that provided several measures of behavior. Off-line, raw force–time data were parsed into individual waveforms (identified by computer as response bouts consisting of at least 4.36 s of dipper-up responding). Individual waveforms of responses (i.e., a response was one drinking bout of about 6 s) were ensemble averaged in the time domain for the first 4.36 s of the event. Two measures were derived from these averaged waveforms: the maximum force attained by the rat during the first second (max force), and the mean force sustained during the remaining 3.36 s (hold force). To quantify isometric force-related forelimb tremor, the 3.36-s hold segment of each individual force–time waveform was subjected to a prime factor fast Fourier transformation (Alligator Technologies, Fountain Valley, CA) to yield separate power spectra for each event. The spectra were then ensemble averaged in the frequency domain for each subject. Integrated power in the 10- to 25-Hz frequency band was taken as the sum (i.e., integral) of the frequency estimates lying within the 10–25 Hz frequency band, and was the area under the curve (power within a given bandwidth) when examining power spectra. Lower frequencies were not included in the tremor measurement because previous analyses showed that frequencies in the region of 7 Hz are the results of the rat’s lapping behavior (rhythmic tongue movements) during the task, and the aim of the research was to quantify forelimb tremor, not movements of the tongue. In addition to the force and tremor measures, time on task (TOT), which is analogous to operant response rate and is the cumulative time of operandum contact during the 8-min session measured to the nearest second, was also cal-
288
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
culated for each rat. TOT was defined in terms of a 1-g forelimb force threshold (i.e., any touching of the operandum counted even if it did not produce water).
Table 1 Means (bold type) and standard errors of the mean (below means) for the indicated dependent variables for the first and last days of exposure to the three different force requirements during the PF ⬎ HF manipulations
2.5. Quantitative analysis For statistical analyses, data for each dependent variable were taken from the initial and final days of each manipulation. Data were analyzed using a two-way, repeated-measures factorial analysis of variance (ANOVA) using SYSTAT’s Multivariate General Linear Model. One factor was the “practice-effects” factor (first versus last day of each phase), and the other was the factor for the required force condition. Data for the PF ⬎ HF manipulations, the PF ⫽ HF manipulations, and the DAFR manipulation were analyzed separately. To quantify the effect of practice on the relationship between isometric force output and tremor across the force requirement manipulations, linear regression was used; log power was regressed on mean hold force separately for the first and last days of performance for each of the parameter values of the PF ⬎ HF and PF ⫽ HF conditions. 3. Results 3.1. PF ⬎ HF conditions Means and standard errors of means for dependent variables measured during this condition are presented in Table 1. Figure 1 illustrates the effects of increasing the peakforce requirement on the shape of one rat’s raw force–time waveforms, and the top panel of Fig. 2 (which includes averaged force–time waveforms for all three PF ⬎ HF conditions) illustrates these effects on the averaged waveforms for all rats (see Figs. 1 and 2, top panel). As expected, max force increased as the peak-force criteria were increased, F(2, 44) ⫽ 144.775, p ⬍ 0.001; this measure did not significantly change from the first day to the final day of the phases. Although rats were not required to increase their dipper-maintaining hold forces during this phase, the hold portion of their responses also increased with increases in the dipper-raising force requirement (see Fig. 2, top panel). This increase in hold force as a function of dipper-raising force criterion was also significant, F(2, 44) ⫽ 42.239, p ⬍ 0.001. Like max force, hold force did not significantly change between the first and last days (see Table 1). Power spectral analyses revealed that integrated power across almost the entire 0–25-Hz frequency range increased with increased force output (see Fig. 2, bottom panel). This was confirmed for the integrated power in the 10–25-Hz band, as the measure increased significantly with the increased force requirements, F(2, 44) ⫽ 31.061, p ⬍ 0.001 (see Table 1). There was a significant practice effect for integrated power in the 10–25-Hz band, as (with the exception of the 60.0/6.7 condition) the measure decreased from the first day to the final day of each 2-week practice period, F(1, 22) ⫽ 6.432, p ⬍ 0.05. With respect to task engage-
20.0/6.7
40.0/6.7
Variable
First
TOT (s)
285.32 289.56 279.20 306.60 300.23 307.35 8.18 4.97 13.15 8.33 10.08 9.53 31.96 32.58 40.92 41.54 54.60 55.58 1.41 1.59 1.07 1.14 1.25 1.14 24.73 25.79 28.38 28.99 31.53 33.24 1.26 1.33 1.60 1.49 1.51 1.52
Max Forcea,b (g) Hold Forceb (g) Integrated power in the 10–25 Hz bandb,c (log10 power)
4.66 0.14
Last
4.46 0.16
First
60.0/6.7
4.87 0.15
Last
4.72 0.13
First
5.09 0.13
Last
5.08 0.11
a
The values for max force are taken from the averaged force-time waveforms. Due to variation in latencies to reach peak forces in individual responses, the values for this dependent variable were lower than the actual peak forces emitted. b Significant force-requirement effect, F-test, df ⫽ 2,44, p ⬍ 0.01. c Significant practice effect, F-test, df ⫽ 1,22, p ⬍ 0.05.
ment as measured by TOT, neither force requirement nor practice significantly influenced the measure during any of the PF ⬎ HF manipulations. 3.2. PF ⫽ HF conditions Means and standard errors of means for dependent variables measured during this condition are presented in Table 2. Due to nonresponding by six rats on the final day of the 60.0/60.0 phase, 18 rats were included in the ANOVA equations during the PF ⫽ HF conditions for all dependent variables except for the time-on-task measures (where nonresponders produced a value of 0). Figure 1 illustrates the way in which increasing the PF ⫽ HF requirements influenced the shape of one rat’s raw force–time waveforms. The averaged force–time waveforms depicted in the top panel of Fig. 3 shows the effect of the force requirements on the force– time waveforms of the entire group. Increases in the peak force and hold force requirements produced significant increases in both max force, F(2, 34) ⫽ 174.412, p ⬍ 0.001 and hold force, F(2, 34) ⫽ 153.066, p ⬍ 0.001. In addition, there was a significant practice effect for hold force, as the measure generally increased from the beginning to the end of each practice period, F(1, 17) ⫽ 23.906, p ⬍ 0.001. A significant practice-by-force requirement interaction affirmed that the increases in hold force as a function of practice were greater as the force requirements were increased, F(2, 34) ⫽ 5.382, p ⫽ 0.009. Max force did not significantly change with practice (i.e., first-day vesus last-day comparison). Integrated power in the 10–25-Hz band also increased as a function of increases in peak force and hold force requirements, F(2, 34) ⫽ 77.910, p ⬍ 0.001. This variable did not significantly change within the PF ⫽ HF 2-week periods. Generally, the PF ⫽ HF manipulations produced an in-
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
289
crease in TOT from the 20.0/20.0 phase to the 40.0/40.0 phase, then a decrease when the contingencies were increased to 60.0/60.0, which is supported by a significant practice-by-force-requirement interaction, F(2, 46) ⫽ 3.236, p ⬍ 0.05. 3.3. DAFR Condition Means and standard errors of means for dependent variables measured during this condition are presented in Table 3. Comparison of the averaged force–time waveforms for the 20.0/6.7 condition in Fig. 4 with Fig. 2 reveals that rats’ force output during this later occurring DAFR condition had increased since the PF ⬎ HF conditions. Inspection of Table 3 suggests that the force increase was a result of training prior to this phase and not the alternating force requirements during this phase, because the increase was present at the beginning of the DAFR condition, and there was no practice effect during this condition. Notwithstanding the baseline increase in force output for the 20.0/6.7 conditions, ANOVAs revealed significant required force effects for max force, F(1, 22) ⫽ 36.999, p ⬍ 0.001; and hold force, F(1, 22) ⫽ 24.17, p ⬍ 0.001 during the daily alternation procedure (see Fig. 4, top panel). In addition to increases in max force and hold force, ANOVAs revealed a significant increase in integrated power in the 10–25-Hz band as a function of force requirement, F(1, 22) ⫽ 24.383, p ⬍ 0.001 (see Table 3). Although TOT did not vary significantly as a function of required force, the measure did increase throughout the manipulation, yielding a significant practice effect, F(1, 23) ⫽ 16.37, p ⫽ 0.001. 3.4. Effects of practice on force-related tremor The effects of practice on the relationship between isometric force output and tremor are illustrated in Fig. 5. The regression lines for the first and last days of practice during each force-requirement phase are highly log linear and approximately parallel, with the tremor power on the last day falling consistently below the tremor power for the first day of data for each force requirement, even though the hold force rises somewhat (e.g., compare the circle and square at the upper right of both functions) during the 2 weeks of
Fig. 1. Force–time waveforms from a representative rat during portions of the final sessions of each of three types of force requirement conditions: (A) peak force greater than hold force; (B) peak force equal to hold force; (C) daily alternation of force requirement. For each condition, time proceeds from the left to the right of each row and from the bottom row to the top row. Force is plotted on the y-axis, where the vertical line on the left end of the bottom row of each graph represents 60.0 g. For each condition, as the lines get darker, higher force requirements are denoted (see legends). This rat was already engaged in the task with the forelimb maintaining downward force on the operandum and drinking water when the plot began. Note the regular force oscillations during the relatively flat (i.e., zero slope ) portions of the records.
290
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293 Table 2 Means (bold type) and standard errors of the mean (below means) for the indicated dependent variables for the first and last days of exposure to the three force requirements during the PF ⫽ HF manipulations 20.0/20.0
40.0/40.0
60.0/60.0
Variable
First
First
First
TOTc (s)
256.21 265.09 254.65 285.99 253.66 227.60 11.65 10.30 14.48 15.49 22.88 27.59 42.33 42.18 59.39 60.12 72.45 77.46 1.86 1.91 1.25 2.26 1.65 2.27 33.55 34.97 49.27 52.48 61.81 70.85 1.39 1.40 1.58 2.15 3.06 2.74
Max Forcea,b (g) Hold Forceb,c,d (g) Integrated power in the 10–25 Hz bandb (log10 power)
5.14 0.10
Last
4.92 0.13
5.94 0.11
Last
5.73 0.12
6.48 0.15
Last
6.56 0.08
The values for max force are taken from the averaged force-time waveforms. Due to variation in latencies to reach peak forces in individual responses, the values for this dependent variable were lower than the actual peak forces emitted. a Significant force-requirement effect, F-test, df ⫽ 2,34, p ⬍ 0.01. b Significant practice effect, F-test, df ⫽ 1,17, p ⬍ 0.01. c Significant practice-by-force requirement interaction, F-test, df ⫽ 2,46, p ⬍ 0.05. d Significant practice-by-force requirement interaction, F-test, df ⫽ 2,34, p ⬍ 0.01.
4. Summary
Fig. 2. Averaged force–time waveforms (top panel) and power spectra (bottom panel) for all rats during the final sessions of the PF ⬎ HF manipulations. Note that the peaks for the averaged force–time waveforms for the 40.0/6.7 and 60.0/6.7 conditions are lower than the required peak force criteria. This is a result of averaging waveforms with varying latencies to reach the actual peak in each individual response.
To summarize, rats were able to emit the forelimb forces that were required during the PF ⬎ HF and PF ⫽ HF manipulations. Forelimb forces generally increased from the beginning to the end of each manipulation (i.e., with practice). Increased force requirements did not significantly decrease rats’ engagement in the task. Indeed, task engagement measures increased with practice during each phase. As isometric force output increased, forelimb tremor increased concurrently. However, with the exception of the highest force conditions (i.e., 60.0/6.7 and 60.0/60.0), forelimb tremor power decreased with practice. 5. Discussion
training. In each of the six cases (three PF ⬎ HF and three PF ⫽ HF conditions) the ratio of tremor power to mean hold force was lower on the last day compared to the first day (p ⬍ 0.05, Wilcoxon rank sum test). This result indicates that tremor power decreased with practice (or physical training) when the level of the hold force was held constant. Note the rightward shift of the squares versus the circles in Fig. 5. This shift establishes that, while increased tremor accompanied increased isometric force output, with practice, higher forces were attained without increases in tremor. Inspection of the first and last days’ values for the daily alternation data (triangles) suggests that physical training effects did not continue in this condition, probably because substantially higher force requirements had preceded this phase of the work.
We report here increases in integrated power in the 10– 25-Hz frequency band (i.e., isometric muscle tremor) as a result of increases in voluntary force output in rats. Previous studies with healthy human subjects have demonstrated that increases in isometric force output elevate power in the frequency range that was elevated in the present study [3–6]. In an earlier review, it was argued that the spectral analysis of force records provides valuable information about not only increases in tremor, but also about the pooled activity of motor units as a result of synchronization and recruitment [1]. Moreover, it has also been suggested that the spectral analysis of muscle force activity provides information regarding asynchronous firing of muscle units (and, therefore, recruitment) that is superior to information provided by surface EMG analysis, because the higher frequencies of the
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
291
Table 3 Means (bold type) and standard errors of the mean (below means) for the indicated dependent variables for the first and last days of exposure to the condition involving daily alternation of force requirements 20/6.7
60/20
Variable
First
Last
First
Last
TOTc (s)
303.94 7.119 47.208 2.66 39.195 2.527
331.759 7.673 47.401 2.757 39.828 2.734
290.543 12.557 64.396 1.895 50.69 2.18
333.463 10.885 62.472 1.836 51.73 2.395
5.097 0.163
5.095 0.182
5.775 0.103
5.742 0.106
Max forcea,b (g) Hold forceb (g) Integrated power in the 10–25 Hz Bandb (log10 power) a
The values for Max Force are taken from the averaged force-time waveforms. Due to variation in latencies to reach peak forces in individual responses, the values for this dependent variable were lower than the actual peak forces emitted. b Significant force requirement effect, F-test, df ⫽ 1,22, p ⬍ 0.01. c Significant practice effect, F-test, df ⫽ 1,23, p ⬍ 0.05.
Fig. 3. Averaged force–time waveforms (top panel) and averaged power spectra (bottom panel) for all rats during the final sessions of the PF ⫽ HF manipulations. See Fig. 2 caption for further explanation.
EMG spectrum may be confounded by covarying electrical phenomena such as the conduction velocities of muscle fibers [4]. Indeed, the EMG power spectrum is influenced by not only muscle fiber conduction velocities, but also by the motor unit action potential, the number of motor units firing in proximity to the electrode, and the recording conditions [13]. Therefore, because of the number of variables that affect EMG, and because of the reported correlation between the results produced through these two methods [5], for motor unit phenomena ranging between 1 and 25 Hz, the spectral analysis of isometric muscle force variation may be a desirable and efficient alternative to more invasive methods (e.g., the use of intramuscular or surface electrodes or accelerometers). We have determined the peak in the power spectrum around 7 Hz to be attributable to rats’ licking frequencies (unpublished observations; force exerted by the tongue on the water cup was measured simultaneously with the forelimb force). In addition to this spectral peak, there are two
other less prominent peaks in the spectra—one at around 13 Hz, and one at around 20 Hz (e.g., see bottom axes in Figs. 2, 3, and 4). In a study examining frequency peaks during human finger muscle contraction, one group reported prominent peaks at 10, 20, and 40 Hz, and attributed their regularity to rhythmic output from the central nervous system [5]. This group argued that, because of the regularity of these peaks under varying force conditions, a central oscillator may be functioning to provide “timing” signals in the form of frequency-coded muscle commands. This argument is consistent with current theories regarding the role of rhythmic activity in the cerebellum in coordinating cell ensembles as they govern skeletal muscles [14]. The fact that there was a peak at 20 Hz in the present study suggests that whatever underlies this peak in humans may also produce this peak in rats. Similarly, although the peak around 13 Hz does not exactly match the 10-Hz peak reported in humans, it could also be an example of a rhythm generator conserved across species, only in the present case it could have been influenced (i.e., shifted rightward) by the prominent peak around 7 Hz. This explanation raises the issue of whether the 13 and 20 Hz peaks observed in this study represent harmonics of the dominant 7-Hz peak. Although we cannot entirely rule out this possibility, previous data from our laboratory support the relative independence of these peaks. In a study in which low doses of the tremorogenic agent harmaline was administered to rats in this procedure, we reported substantial increases in the power of the frequency spectrum around 13 Hz in the absence of increases in the height of the peak around 7 Hz [15]. Interestingly, however, the peaks associated will the 7-, 13-, and 20-Hz regions of the power spectrum were all shifted leftward by harmaline. Therefore, determination of the degree to which the phenomena reported in this study are analogous to human studies awaits
292
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
Fig. 5. Group mean integrated power in the 10–25 Hz band (tremor) as a function of mean hold force during the PF ⬎ HF, PF ⫽ HF, and DAFR manipulations for the first and last days of experience with each of the different force requirements. Regression lines show the relationship between isometric force output and tremor early (the first day, upper line) and late (last day, lower line) in training with the PF ⬎ HF and PF ⫽ HF manipulations (data for the DAFR conditions were not included in the regression line calculations). Note that higher isometric forces were attained with less forelimb tremor after 2 weeks of training, except in the alternation procedure.
Fig. 4. Averaged force–time waveforms (top panel) and power spectra (bottom panel) for all rats during the final sessions of the procedure involving daily alternation of the force requirements (DAFR). See Fig. 2 caption for further explanation.
the analysis of force recordings in rats that do not contain the 7-Hz peak (i.e., in a pressing-without-licking preparation). The fact that force-related increases in the power spectrum were not limited to the frequencies between 10 and 25 Hz, especially in the PF ⫽ HF condition, raises the question of whether the increases in power observed here are merely attributable to the increased amplitude of the force signals (i.e., instead of changes in the behavior of motor units). Although we cannot rule this possibility out, our demonstration that, with practice, increased force output was accompanied by decreased tremor, suggests otherwise. Moreover, we have previously reported tremorogenic drug-induced increases in the heights of various peaks of the power spectra in the absence of increases in hold force [10,15]. Taken together, these findings support our hypothesis that the phenomena measured in this study were specific to isometric force-related changes in motor unit ensemble activity.
In the present study, there was an overall tendency for practice to affect the relationship between isometric forelimb force output and tremor. This effect was manifested by the generation of higher forces while tremor generally decreased. Although we expected the force–tremor relationship to be influenced by practice, previous studies examining the effects of isometric training on EMG measures have demonstrated training-induced increases in motor unit activity [16]. Two physiological adaptations have been proposed to contribute to training-related increases in force output: muscle hypertrophy and neural adaptation (in the form of the recruitment of a “functional reserve” of previously inactive motor units with training) [16]. Although alterations in motor unit activity likely account for the observed force requirement effects on tremor, our results suggest that the ability of our rats to emit higher forces with diminished tremor over time may have resulted from training-induced muscle hypertrophy. The results in the present study are based on forces generated by rats under voluntary, sustained force output conditions in which a downward press with a single designated forelimb was required. Rats are capable of producing forces in excess of those required in this study by emitting either ballistic operandum striking behaviors (e.g., [11,17]) or rapid grasp-and-pull behaviors (e.g., [18]). Although these behaviors can yield valuable information, there are several reasons why sustained, nonballistic responses such as those used in the present study are valuable. The first reason is
J.A. Stanford et al. / Physiology & Behavior 69 (2000) 285–293
that, to spectrally analyze force recordings so that biophysical phenomena ranging from low frequency (e.g., licking) to high frequency (e.g., muscle unit activity) can be assessed, sustained force output (i.e., output enduring for several seconds) is required to provide adequately long records for meaningful Fourier analysis. In the present case, the power spectrum ranges from 0 to 25 Hz. This means that, according to a common rule of thumb stating that frequency resolution is generally the reciprocal of the observed time interval [19], time samples of at least 1 s must be recorded (and, indeed, samples of greater durations than the minimum based on this rule allow for increased accuracy of estimation and reliability). Samples longer than 1 s cannot be accomplished with a rat’s ballistic responses. Another rationale for utilizing sustained isometric forelimb force is that, although tremor and other neuromuscular phenomena can be measured by using nonisometric responses (e.g., [5]), the signals recorded from moving muscles contain frequencies that result from mechanical properties of these movements [4,5]. Regarding force differentiation as measured in the daily alteration procedure (DAFR), the fact that rats were generally able to produce appropriate force output as evidenced by the significant main effects for force requirement during this phase, suggests that rats were able to discriminate the force requirements and adjust their behavior accordingly when the differences between the requirements were suitably large (20.0/6.7 versus 60.0/20.0). In a study examining the effects of different response topographies on force discrimination in rats, we previously demonstrated that rats were better able to confine their responses to force bands (lower and upper force requirements) when the required response was a downward press than when the response was a grasp and pull in the horizontal plane [18]. We concluded that the grasp-and-pull topography may have elicited unconditioned species-typical feeding, burrowing, or scratching tendencies that interfered with the conditioning of circumscribed force output. Because the present contingencies did not include force “bands” (i.e., there was no upper force limit), the present findings suggest that rats tended to minimize response cost by somewhat limiting their force output during the low-force conditions. The results revealed interesting information regarding the effects of varying force requirements on fine motor behavior in rats. The measurement of voluntary force output, as well as high-frequency variations in muscle force (i.e., tremor) as a result of voluntary force output, from rats provides a homologue for similar studies being conducted with humans [4,5]. The use of the forelimb tremor task procedure with rats could provide valuable, noninvasive, functional assays for nonhuman models of peripheral muscular and neuromuscular disorders, such as the 6-OHDA–induced Parkinsonism model or the sciatic nerve constriction neuropathy model. Furthermore, creative programming and utilization of concurrent task engagement and force mea-
293
sures may also allow for cross-species usage of terms such as maximal voluntary force (and percentages thereof) in studies assessing isometric training in physical therapeutic contexts.
Acknowledgments This research was supported by MH43429.
References [1] Freund H-J. Motor unit and muscle activity in voluntary motor control. Physiol Rev 1983; 63:387–436. [2] Fuglevand AJ. Neural aspects of fatigue. Neuroscientist 1996;2:203–6. [3] Allum JHJ, Dietz V, Freund H-J. Neuronal mechanisms underlying physiological tremor. J Neurophysiol 1978;41:557–71. [4] Homberg V, Reiners K, Hefter H, Freund H-J. The muscle activity spectrum: spectral analysis of muscle force as an estimator of overall motor unit activity. Electroencephalogr Clin Neurophysiol 1986;63:209–22. [5] McAuley JH, Rothwell JC, Marsden CD. Frequency peaks of tremor, muscle vibration and electromyographic activity at 10 Hz, 20 Hz and 40 Hz during human finger muscle contraction may reflect rhythmicities of central neural firing. Exp Brain Res 1997;114:525–41. [6] Orizio C, Perini R, Diemont B, Figini MM, Veicsteinas A. Spectral analysis of muscular sound during isometric contraction of bicepts brachii. J Appl Physiol 1990;68:508–12. [7] Newland MC. Operant behavior and the measurement of motor dysfunction. In: Weiss B, O’Donoghue J, editors. Neurobehavioral Toxicity: Analysis and Interpretation. New York: Raven Press, 1995. pp. 273–97. [8] Fowler SC, Davison KH, Stanford JA. Unlike haloperidol, clozapine slows and dampens rats’ forelimb force oscillations and decreases force output in a press-while-licking behavioral task. Psychopharmacology (Berlin) 1994;116:19–25. [9] Fowler SC, Liao R-M, Skjoldager P. A new rodent model for neuroleptic-induced pseudo-parkinsonism: low doses of haloperidol increase forelimb tremor in the rat. Behav Neurosci 1990;104:449–56. [10] Stanford JA, Fowler SC. Scopolamine reversal of tremor produced by low doses of physostigmine in rats: evidence for a cholinergic mechanism. Neurosci Lett 1997;225:157–60. [11] Alling K, Poling A. The effects of differing response-force requirements on fixed-ratio responding of rats. J Exp Anal Behav 1995;63:331–46. [12] Makhay M, Alling K, Poling A. Effects of cocaine on fixed-ratio responding of rats: modulation by required response force. Pharmacol Biochem Behav 1994;48:511–4. [13] Fuglsang–Frederiksen A, Ronager J. The motor unit firing rate and the power spectrum of EMG in humans. Electroencephalogr Clin Neurophys 1988;70:68–72. [14] Welsh JP, Lang EJ, Sugihara I, Llinas R. Dynamic organization of motor control within the olivocerebellar system. Nature 1995;374:453–7. [15] Stanford JA, Fowler SC. At low doses, harmaline increases forelimb tremor in the rat. Neurosci Lett 1998;241:41–4. [16] Bandy WD, Hanten WP. Changes in torque and electromyographic activity of the quadriceps femoris muscles following isometric training. Phys Ther 1993;73:455–67. [17] Elsner J, Fellmann C, Zbinden G. Response force titration for the assessment of the neuromuscular toxicity of 2,5-Hexanedione in rats. Neurotoxicol Teratol 1988;10:3–13. [18] Kirkpatrick MA, Fowler SC. Operant force-band differentiation by rats using two different response topographies. Bull Psychonom Soc 1989;27:52–4. [19] Marple SL, Jr. Digital Spectral Analysis With Applications. Englewood Cliffs, NJ: PTR Prentice Hall, 1987.