A new animal model of posttraumatic stress disorder (PTSD) is proposed in which non-habituation of the acoustic startle response is used to identify animal subjects with altered alarm responses subsequent to trauma exposure. Detailed examples of the modeling system are presented that serve to distinguish this model from more commonly used fear-conditioning methodologies.
The psychophysiology of the acoustic startle reflex has been studied for decades in animals and humans. In distinction to the orienting response, which consists of a turning in the direction of a novel sound, the acoustic startle reflex is a survival mechanism of alarm that rapidly alerts and arouses an animal perceiving a loud noise. The acoustic startle mechanism in humans can be routinely measured through muscle movement, and changes in heart rate, and skin conductance.
The acoustic startle reflex in humans and animals has been shown to accommodate to the repeated administration of loud acoustic stimuli, such that the subject no longer startles (startle habituation; Shalev, 1992b). Similarly the acoustic startle response will accommodate to a softer pre-startle warning tone (prepulse modulation; Ornitz, 1989). PTSD is a psychiatric condition characterized by three clusters of symptoms that arise in the aftermath of a traumatic event. One cluster is composed of reexperiencing phenomena, including distressing memories of the event; the second cluster contains symptoms of avoidance of reminders of the event, and of emotional and cognitive numbing; and the third cluster includes symptoms of psychophysiological arousal (American Psychiatric Association, 1994).
PTSD patients have been shown to demonstrate autonomic arousal in response to audiovisual presentations of traumatic scenes (Pitman, 1987; Shalev, 1992a). A salient feature of the disorder is the absence of extinction of a symptomatic response to traumatic cues. Further, a commonly reported diagnostic criterion is a reflex exaggerated startle response to acoustic stimuli.
The acoustic startle reflex has been studied in subjects suffering from a variety of psychiatric disorders. Of particular interest is the finding that the normal capacity for both prepulse modulation and habituation of the acoustic startle reflex is disrupted in PTSD. Shalev has shown that patients who developed PTSD from various types of trauma fail to inhibit their acoustic startle reflex, which may be related to the exaggerated startle response. Importantly, patients with other anxiety disorders show the usual pattern of startle habituation. Thus, the absence of startle habituation appears to be a specific biological marker of PTSD (Shalev, 1992b).
We are reporting a major advance in the development of an animal model of PTSD. We have chosen to use rats as the animal, electric tail shock administered in a confined space as the trauma, and acoustic startle habituation as the diagnostic marker of PTSD. In this publication we are reporting pilot data chosen as illustrative of the groups of habituators, non-habituators and controls. The full data set will be published at a later date.
Subjects were mycoplasma free male Sprague Dawley rats weighing approximately 275 g (Harlan, Indianapolis, IN). They were housed in groups and allowed ad libitum food and water (except while being tested) in 25 degrees C. room on a 12:12 light:dark cycle for at least two weeks prior to experimental procedures. Institutional Review Board approval was obtained for the performance of these procedures.
The basic startle apparatus has been described (Leitner, 1986). Briefly the subjects are placed in a plastic box measuring (13"x 7.25"x 4.25") which sits on 4 piezoelectric transducers (Edmund Scientific, Barrington, NJ) located at each of the four corners. The box was surrounded by side rail guides to limit lateral movement, !" wider than the box at each side. Movement by the rat, including startle, caused the piezoelectric transducers to output a voltage with a linear relationship to the force applied to the floor of the cage. This system was housed inside a ventilated, lighted, sound attenuating chamber. One piezoelectric wide-dispersion horn (Radio Shack #40-1379) was mounted directly over the cage.
The voltage from the cage transducers was fed into an amplifier and then into an analogue to digital module and a PC computer at a rate of 50 hz/channel(#S77-08, #L19-01, High Speed Videograph, Codas Software;Coulbourn Instruments, Allentown, PA) . The signals were subsequently analyzed using analysis software (Dadisp, DSP Corporation, Cambridge, MA) programmed to quantify the peak amplitude and peak latency occurring subsequent to startle stimuli.
The horn delivering startle stimuli was driven by a power amplifier and a white noise generator (S81-02, Coulbourn Instruments). The signals were programmed using behavioral control software (L2T2, Coulbourn Instruments). The signals controlling the white noise generator were also directed to a second channel of the A/D converter input in order to document the coordination of the stimulus with the startle reaction.
Intermittent tail shock was performed as described previously (Garrick, 1989). The rat was loosely restrained in a plexiglas tube inside a sound attenuated, ventilated, illuminated box. Electrodes were taped to the tail which extended out the end of the tube. A constant current shock generator (Lafayette Instrument Co., Lafayette, Ind.) provided intermittent shocks (5 sec 1 ma unpredictable tail shocks, average 1/min) controlled by a Commodore microcomputer.
In the first phase of the experiment, each subject was tested for baseline startle reactivity. Each subject was placed in the startle measurement apparatus where it was presented with a series of 150 startle-eliciting bursts of white noise (intensity 125 dB SPL; 0 msec rise-fall time; duration 10 msec) at 6 sec intervals. Data was collected as noted above. Only startle stimuli number 5-150 were considered for analysis of habituation, since the study animals demonstrated variable initial startle reactions and habituation to repeated acoustic presentations was the focus of the study.
In the second phase of the experiment, the experimental subjects (n=2) were exposed to a three day course of traumatic shock intended to elicit a PTSD-like state. On day one, the subject received 30 tail shocks with a variable interval (VI), average 1/min, schedule, and then returned to their home cage. On day two, after being placed in the same plexiglas cage for 8 min, the subjects received 5 additional tail shocks VI 5.6 min average. On the third day, the subjects were placed in the shock chamber without any electric shock. A control animal was placed through the same procedures, without any electric shock delivery.
In the third phase of the experiment subjects were tested for startle reactivity over time. This testing, as described above, was performed prior to experimental (or control) procedures (Day 0) and then for 6 days in the 8 days following termination of startle (day 1, 4,5,6,7,8, skipping weekend days) and then weekly thereafter for 1 month.
Three subjects, one control and two experimental, are presented in detail. These subjects are selected since they demonstrate representative patterns of startle response. Subjects variably displayed in the first few startle presentations, reactions exceeding 0.5 v. These early startle reactions were therefore not included in the average amplitudes analyzed. The control subject demonstrated minimal change in startle amplitude over time, with average amplitude being 0.18 + SD .07 Volts for startle numbers 5-150. The average startle amplitude on day 0 was .16 V, and day 1 was .25 V. Three weeks after the procedure, the average startle amplitude was .14 V.
The two experimental subjects demonstrated divergent responses. The first experimental subject demonstrated a pattern resembling the control subject. Its average startle amplitude over all startle testing days was .30 + SD .05 V as compared to the average startle prior to stress exposure (Day 0) which was .34 V, and on the day immediately following stress exposure (Day 1) of .18 V. There was some variability in the post-stressor amplitudes with an increase only on day 4 post stress (.48 V), other test days being considerably lower than average. Three weeks after stress exposure, the startle response measured .30 V.
The second experimental subject demonstrated a significant nonhabituation of startle reactivity, with an average amplitude over all testing days .46 + .1. On Day 0, this subject's startle reactivity was .34 V, rising to 0.47 V the first day after stress exposure. It continued high with the exception of two days (Days 3 and 4) when it was the same as pre-stress level. The startle reactivity continued high during the weeks 2nd and 3rd week post stress (.67 v and .48 V respectively)
Startle latency was equivalent in all three subjects in that 70- 80% of all startle peaks occurred in less than 100 msec following onset of startle stimulus.
Using a prolonged session of startle-eliciting acoustic stimuli the control rat rapidly and consistently habituated without persistently high startle reactions. Subjects stressed with intermittent tail shock over a brief period of time demonstrated two patterns. One subject, resembling the control subject, consists of rapid habituation to the startle in the first session and a similar response in subsequent startle trials. The second subject demonstrated a pattern akin to the non-habituation described in PTSD patients (Shalev, 1992b), in which the animal failed to habituate to startle stimuli. This was particularly striking since the subject was exposed to a large number of startle eliciting stimuli over several trials over several weeks duration. The two experimental subjects initially had a somewhat higher than average startle reaction when tested prior to stress exposure. However, since both subjects demonstrated similar initial startle reactivity, this alone cannot account for the later non-habituation of one.
The uniqueness of this animal modeling system comes from its focus on non-habituation of startle in the absence of traumarelated cues. It is the continued hyperarousal in the absence of specific cues, typified by non-habituation of the startle reactions (Shalev, 1992b) that more accurately reflects the clinical situation. This is not a fear conditioning or inescapable shock (IES) paradigm as used previously to model PTSD (Ottenweiler, 1989; Davis, 1984). The IES model is a model of experimental depression, usually using foot shock and a shuttle box learning paradigm. The inescapable nature of the shock is not dissimilar, but our protocol is not paired to a learning task and contains startle reactivity which has not been a feature of IES. The testing situation does not use a fear-potentiated startle reaction (Davis, 1978). The subjects are in a small plastic restraining cage with electrodes taped to their tail in the induction phase and are freely moving in a plastic box with a wire top for startle testing. There may be a minor element of fear potentiation since stress and startle are performed in the same laboratory with the same sound-attenuated chamber.
We are reporting a small sample of our experimental subjects to illustrate the feasibility and validity of this animal model. It is unclear at this time whether changes in the stress paradigm with respect to shock intensity, volume, or timing will affect the number of subjects developing non-habituation of their startle reactions. Nor is it clear that repeat exposure to the chamber without experiencing tail shock affects the number of subjects developing non-habituation of startle. The second session of tail shocks appears to be critical; an earlier unpublished pilot study (n=40) in which subjects received 60 intermittent tail shocks was unable to produce any subjects that developed startle non-habituation. It is currently unclear, therefore, which parameter is most relevant for future studies.
It is also unclear whether the methods used in this study actually minimize the degree of startle, as the subject is preexposed to the startle testing chamber prior to receiving stress exposure and receiving a large number of startle presentations on a fixed ratio schedule. Potentially these factors could inhibit subsequent startle reactivity by familiarizing the subject with the testing situation. It is noteworthy that as a result, this design departs from a faithful modeling of the clinical situation in which PTSD patients have not been familiarized with testing situations prior to experiencing their trauma. A cross sectional design with appropriate control subjects would be helpful to validate this model. However, that design would not address the nature-nurture questions at the center of this research effort.
A small subset of subjects exposed to stress exposure develop non-habituation of startle reactions. This fact is in accord with human studies. It is well known that not all persons exposed to traumatic stimuli develop a post-traumatic stress syndrome. The National Vietnam Veterans Readjustment Study(NVVRS) documented that about 30% of a large cohort of Vietnam combat veterans could be diagnosed as having PTSD at some point since the war. Shore found that 11% of residents living near the Mt St Helen's volcano endorsed symptoms of PTSD(Shore, 1986). Thus, across different types of events, about 11-30% of persons exposed to traumas develop PTSD as a function of intensity and chronicity. That a consistent subset of samples develop this disorder may be a consequence of biological factors independent of the trauma itself. The current study sample documented three subjects which all responded to startle stimuli similarly prior to stress exposure; namely, by reliable habituation of startle reactivity. Following the stress exposure, however, only one of the two subjects in the paradigm continued to accommodate to startleeliciting stimuli; the second demonstrated the non-extinction which seems characteristic of PTSD in humans. A larger cohort is being investigated to determine the precise fraction of subjects which respond similarly to these animals.
The successful development of an animal model of PTSD holds the promise of fostering substantial progress in our understanding of the pathophysiology of PTSD and in investigating possible biological interventions. This model may provide a useful tool for the investigation of specific genetic variables which may predispose to the development of PTSD, as well as associated psychosomatic and stress related physical syndromes. We are reporting an advance, that may herald a new approach to the study of PTSD through basic science research.
American Psychiatric Association(1994). Diagnostic and Statistical Manual IV. Washington DC: American Psychiatric Press, Inc.
Charney DS, Deutch AY, et. al(1993). Psychobiologic mechanisms of posttraumatic stress disorder. Archives of General Psychiatry 50:294-305.
Davis M and Astrachan DI (1978). Conditioned fear and startle magnitude: Effects of different footshock or backshock intensities used in training. Journal of Exper. Psychol. Anim. Behav. Process. 4:95-103.
Davis, (1984). The mammalian startle response. In, Eaton, RC. Neural mechanisms of startle behavior. New York: Plenum Press.
Garrick T, Buack S et al (1989). Unsignalled shock stimulates gastric contractility and causes mucosal injury. Behavioral Neuroscience 103:124-130.
Kulka RA, Schlenger WE et al (1990). Trauma and the Vietnam War Generation: Report of Findings from the National Vietnam Veterans Readjustment Study. New York: Brunner/Mazel.
Leitner, D (1986) Alterations in other sensory modalities accompanying stress analgesia as measured by startle reflex modification. Annals of New York Academy of Sciences 467:82-92.
Ornitz EM and Pynoos RS(1989). Startle modulation in children with posttraumatic stress disorder. American Journal of Psychiatry 146: 866-870.
Ottenweiler JE, Natelson BH et al.(1989). Adrenocortical and behavioral responses to repeated stressors: Toward an animal model of chronic stress and stress-related mental illness. Biological Psychiatry 26:829-841.
Pitman RK, Orr SP, Forgue DF, de Jong JB, Claiborn JM (1987). Psychophysiology of PTSD imagery in Vietnam Combat Vets. Arch Gen Psych 44:970-975.
Shalev AY, Orr SP and Pitman RK (1992a). Psychophysiologic response during script-driven imagery as an outcome measure in posttraumatic stress disorder. Journal of Clinical Psychiatry 53:324-326.
Shalev AY, Orr SP et al(1992b). Physiologic responses to loud tones in Israeli patients with posttraumatic stress disorder. Archives of General Psychiatry 49:870-875.
Shore JH, Tatum EL and Vollmer WM (1986). Psychiatric reactions to disaster: The Mount St. Helens experience. Am J Psychiat 143:590-595.
This work was supported by the Department of Veterans Affairs, Psychiatry and Research Service, Merit Review Research Program.
Copyright 1995 Traumatology Forum
Green Cross
Forum
