top of page

DEEP DIVE into Learned Helplessness: Applications in Animal Behavior




Introduction

“Learned helplessness” is a term first used by Overmier and Seligman (1967) to describe the phenomenon that occurs when an animal has been exposed to an inescapable aversive stimulus and subsequently ceases attempts to avoid or escape the aversive stimulus when it is possible to escape. They hypothesized that animals learn responding and termination of or reduction in the aversive stimulus are independent of one another (Overmier & Seligman, 1967). Seligman (1972) further described learned helplessness as a response to uncontrollable trauma having three characteristics: passivity when exposed to trauma, failure to learn that responding to trauma can provide relief, and a higher stress response compared to the stress response to controllable trauma. Many researchers studied this phenomenon in dogs and rats in the latter half of the 20th century, but the more recent studies on learned helplessness have shifted to using induced learned helplessness in rats and mice as models for human major depressive disorder (MDD) and post-traumatic stress disorder (PTSD) (Conoscenti & Fanselow, 2019). The comparatively vast body of research dedicated to the application of the learned helplessness model to human psychological research is beyond the scope of this review; instead, this review focuses primarily on the studies of learned helplessness in dogs, the proposed mechanisms behind the phenomenon, and contexts in which knowledge of the phenomenon may be applicable in companion animal behavior.


Proposed Mechanisms: Operant Behavior

Two contradictory hypotheses were proposed to explain what was initially termed the interference phenomenon, the latency or failure to respond to aversive stimuli or trauma after being exposed to uncontrollable shock (Overmier and Seligman, 1967). The learned helplessness hypothesis suggested the lack of appropriate response is learned and based on the lack of control the animal has over the outcome (Overmier & Seligman, 1967; Seligman & Groves, 1970). On the other hand, the motor activation deficit hypothesis was proposed by Weiss and Glazer (1975a), who insisted the response could not possibly be learned, but resulted from a temporary physiological change. A review of the studies resulting in these conflicting opinions follows.


Overmier and Seligman (1967)

Overmier and Seligman (1967) ran three experiments to test two existing hypotheses for the interference phenomenon associated with escape/avoidance behaviors after exposure to inescapable shocks, and to test effects of the passage of time and other varied conditions. All three experiments shared some common design components. Subjects were adult mongrel, or mixed-breed, dogs, split into various treatment and control groups. None of the dogs were used in more than one experiment. Treatment groups received one session of inescapable, unpredictable shocks delivered through electrodes secured to their hind feet while restrained in a Pavlovian harness. After 24 hours (with the exception of some groups in Experiment 3), all groups entered a separate testing unit for 10 trials of avoidance training. The avoidance training occurred in a shuttlebox with a “shock” side and a “safe” side separated by a barrier that could be jumped to escape the shocks delivered through the floor on the “shock” side. Results for each group were compared within and across experiments.

            Experiment 1 varied shock density by manipulating the frequency and duration of shock applied to each group. They split dogs into 4 groups of eight. Group 1 was the control, receiving no shock. Groups 2 and 3 were high shock density groups. Group 2 received 64 shocks 5 s in duration every 60-120 s. Group 3 received 640 shocks 0.5 s in duration every 4.5-18 s. Group 4 was the low shock density group, receiving 64 shocks 0.5s in duration every 60-120 s. Results showed a significantly greater failure to escape and latency to respond in all treatment groups compared with the control group, with the high shock density groups showing the greatest interference effect.

            Experiment 2 was designed to test the hypothesis that dogs exposed to inescapable shock learn “instrumental skeletal-motor responses” (p. 28) that are incompatible with escape/avoidance behavior, and the authors cite the following supporters of this idea: Adams and Lewis (1962), Dinsmoor and Campbell (1956), McAllister and McAllister (1962), and Mullin and Morgenson (1963). It also tested the hypothesis that adaptation or habituation to shock leads to a lack of motivation to escape or avoid future shocks, citing MacDonald (1946) as the supporter of this idea. Overmier and Seligman tested the effects of a paralytic agent, curare, and increasing shock intensity on the latency and failure to respond in a similar manner as described in their first experiment. They split dogs into 3 groups of eight: high-motivation, curare-shock, and curare-no shock. The same shock treatment applied to Group 2 (high shock density) in the first experiment was applied to the high-motivation and curare-shock groups, except that the high-motivation group received higher intensity shocks during avoidance training in order to increase motivation to escape. The curare-no shock group was chemically immobilized but not exposed to inescapable shock. They found that Group 2 and the high-motivation group were not significantly different in their latency or failure to respond in avoidance training, and Group 2 and the curare-shock group were not significantly different in their latency to respond. Moreover, Group 1 and the curare-no shock group were not significantly different in their latency to respond, but the two curare-treated groups were significantly different in their latency and failure to respond.

            Experiment 3 examined the effect of the passage of time between inescapable shock and avoidance training on latency and failure to respond to shock. They split dogs into 8 groups of 4 and divided those 8 groups into 2 sets. Set 1 received the same treatment as Group 2 in experiment 1. Set 2 received the same shocks as Set 1 but were also chemically immobilized with curare, given a tone to predict half the shocks, and these tones were counterbalanced with tones that did not predict shocks. Each of the 4 groups in each set experienced a different time delay between treatment and avoidance training, which was 24, 48, 72, or 144 hours. Results showed no significant differences between the groups in the two sets, so the data from the equivalent time-delay groups were combined. All groups that were given avoidance training greater than 24 hours after treatment showed no significant difference from one another or Group 1, but they were significantly different in latency and failure to respond when compared with the 24h groups. Overall, these results suggest an alternative to the proposed hypotheses for the interference phenomenon, which they suggest may be explained by a “massive parasympathetic reaction” (p. 33) or, more likely, a “learned helplessness” (p. 33).


Seligman and Groves (1970)

The previous set of experiments demonstrated learned helplessness was transient: dogs exposed to inescapable or uncontrollable shock that displayed learned helplessness did not sustain this lack of responding 48 hours after exposure. For this study, Seligman and Groves (1970) used the same Pavlovian harness and shuttlebox design as Overmier and Seligman (1967). They evenly distributed 18 lab-raised, unsocialized beagles and 15 adult mongrels with unknown history across three groups: four-spaced, two-spaced, and a control group. The four- and two-spaced groups were subjected to their respective number of days of inescapable, unpredictable shock sessions over the course of 8 days. Each session consisted of 60 shocks lasting 5 s each. Sessions were conducted on days 1, 3, 4, and 8 for the four-spaced group, and days 1 and 8 for the two-spaced group. On day 15, one week after the final shock session, all three groups were subjected to avoidance training. Lights were dimmed to indicate shock was imminent in this study.

Results showed that beagles in both four- and two-spaced groups had the greatest latency and failure to respond, followed by four-spaced mongrels, then two-spaced mongrels, which had similar responding to the control group. The control beagles showed a slightly greater latency and failure to respond than control mongrels. These results contribute two important ideas: they suggest that proactive interference, or previous learning history, may affect retention when learning trauma and responding are independent, and that learned helplessness is not always temporary.


Seligman (1972)

Seligman (1972) succinctly reviewed the research on learned helplessness thus far, describing the behavior patterns observed in the studies. He stated that dogs first exposed to escapable shock responded with progressively shorter latencies as trials progressed, whereas two-thirds of the dogs exposed to inescapable shock before being exposed to escapable shock responded similarly to naïve dogs initially, but then did not continue to respond as such, even if they successfully escaped the shock—they do not learn. Alternative hypotheses for learned helplessness which have been ruled out according to Seligman include adaptation or habituation, sensitization, and incidental or intentional reinforcement or punishment resulting in a motor response pattern that is incompatible with escape behavior. The cure for learned helplessness according to Seligman (1972) is a procedure termed “directive therapy,” during which dogs are dragged from the “shock” side of the shuttlebox to the “safe” side to demonstrate the termination of shock. Prevention of development of learned helplessness by performing behavioral immunization through experience with controllable trauma is also discussed, using rats as an example:

When wild rats are squeezed in the human hand until they stop struggling, then placed in a water tank, they drown suddenly. Unlike nonsqueezed rats which swim for 60 hours before drowning, these rats dive to the bottom and drown within 30 minutes. Sudden death is prevented by a technique which resembles our immunization procedure: If the experimenter holds the rat, then lets it go, holds it again and lets it go, sudden death does not occur. Further, if, after holding it, he puts the rat in the water, takes it out, puts it in again and rescues it again, sudden death is prevented. These procedures, like our own, may provide the rat with a sense of control over trauma and thereby immunize against sudden death caused by inescapable trauma. (Seligman, 1972, pp. 410-411)

This idea is explored further by Ferrandiz and Pardo (1990).


Ferrandiz and Pardo (1990)

Thirty six adolescent to young adult dogs underwent five phases of training to test behavioral immunization to learned helplessness in response to an aversive stimulus. Dogs were split into six groups: controllable and predictable noise-immunized (CP), controllable but unpredictable noise-immunized (CU), uncontrollable but predictable noise-immunized (UP), uncontrollable and unpredictable noise during immunization training (UU), no immunization treatment (NI), and no immunization or inescapable noise treatment (NN). Groups were tested across 5 phases: appetitive contingent training (learning to press a lever to deliver food), immunization training (60 trials, 60 s each), inescapable noise training (60 trials, 60 s each), recovery (3 days), and appetitive noncontingent context (food delivery in training compartments on a random schedule for 10 minutes). In the immunization training phase, CP and CU groups were able to stop an aversive noise by pressing the lever. For predictability, the CP group could see a light turn on just before the noise began. Groups UP and UU were exposed to the same length of noise as CP and CU groups, and NI and NN groups were exposed only to the testing compartment. During the inescapable noise training on the following day, none of the groups received any signaling and the noise sounded on a random schedule. All groups participated except for NN, which was only exposed to the testing compartment. Recovery phase lasted three days, the first of which dogs were left undisturbed in their home cages, and on the second and third days they were exposed to different compartments from the ones they were trained in for an hour each day. Finally, in the appetitive noncontingent phase, all groups received food in the training compartments on a random schedule for 10 minutes. Their results showed that non-immunized animals (UU, NI) responded the least, immunized animals (CP, CU, UP) responded more, and the control group (NN) responded the most. The group that responded most closely to the control group was the controllable and predictable noise-immunized group, suggesting that predictability, not just controllability, of an aversive stimulus may be important in preventing learned helplessness. This finding seems to be corroborated by other researchers (Dess et al., 1990; Minor et al., 1991; and Minor & Hunter, 2002, as cited in Conoscenti & Fanselow, 2019). Moscarello and Hartley (2017) also tend to agree with this finding, building on this research with their ideas of agency, control, and action-outcome learning, which will be described in another section of this review.


Proposed Mechanisms: Physiological Change

Weiss and Glazer (1975a) and Weiss and Glazer (1975b)

            Contrary to the learned helplessness hypothesis, the motor activation deficit hypothesis proposed by Weiss and Glazer (1975a) states that “severely stressful conditions can lead to a deficit in avoidance-escape responding by reducing noradrenergic activity in the brain” (pp. 499), noting also that the altered norepinephrine levels in the brain occur with uncontrollable shock, but not with controllable shock. Weiss and Glazer (1975a) performed their experiments on rats and subjected them to a cold swim as the aversive/traumatic event for most experiments. They provided physiological evidence for their motor activation deficit hypothesis as well as additional evidence for the effectiveness of behavioral immunization through pre-training the appropriate escape/avoidance behavioral response (Weiss & Glazer, 1975a). Weiss and Glazer (1975a) offered an alternative explanation for behavioral immunization, however; rather than attributing to dogs the cognitive ability to recognize and generalize a sense of control, they suggested that response latency is reduced “because the animal, having only a limited ability to mediate motor behavior, does not have to produce a great deal of activity in searching for the correct response but simply has to execute a response it has already learned to be correct” (p. 507). Weiss and Glazer (1975a) ultimately agreed that long term deficits in responding are best explained by a learned response versus a physiological effect, but that their motor activation deficit hypothesis better explained the acute, transient response latency.

            At the end of their second paper, Weiss and Glazer (1975b) directly addressed and compared their methods and results with the Seligman and Groves (1970) study. Both studies used repeated exposure to aversive stimuli but produced significantly different results: Weiss and Glazer (1975b) exposed rats to daily sessions over 15 days and found decreased latency to response as trials progressed, whereas Seligman and Groves (1970) exposed mongrels and beagles to several spaced sessions over 8 days and found the latency or failure to respond was not attenuated one week later. Weiss and Glazer (1975b) suggested that the experiments did not mimic chronic stress in comparable ways and that their study was designed to show that physiologic habituation or adaptation occurs with repeated exposure to aversive stimuli, which it did, and that Seligman and Groves (1970) did not perform treatments in such a way that they would expect the same adaptation to occur. Weiss and Glazer (1975b) also brought attention to the breed differences observed by Seligman and Groves (1970) between beagle responses and mongrel responses, suggesting that the “more robust mongrel animals” (p. 533) were more resistant to nontransient learned helplessness. They do not, however, offer any specific explanation—genetic component, learning history, breed characteristics, etc.—for these differences, and maintain that “the evidence for the learned helplessness hypothesis is not compelling and its utility in animal research is questionable” (Weiss & Glazer, 1975b, p. 533).


Implications and Applications in Companion Animal Behavior

Agency and Control

            Moscarello and Hartley (2017) define agency as an organism’s perception of control over its environment based on its learning history. Further, “agency reflects an estimate, inferred over some aggregate of past experiences, of whether important environmental events and stimuli are predominantly controllable…or are consistently outside one’s control” (Moscarello & Hartley, 2017, p. 1). Seligman and Groves (1970) demonstrated that an animal’s learning history is significant in the acquisition and recovery from learned helplessness, which parallels Moscarello and Hartley’s (2017) proposed relationship between agency and control. In agreement with Ferrandiz and Pardo (1990), Moscarello and Hartley (2017) suggest that action-outcome learning is inhibited by high entropy, or low predictability, leading to more reactive than proactive behavior. Controllability affects defensive behavior: greater control leads to more proactive behavior, whereas less control leads to more reactive or defensive behavior, especially since uncontrollable contexts may not be worth the cognitive and physical effort required to attempt escape or avoidance (Moscarello & Hartley, 2017). Thus, learned helplessness could be a reactive or defensive behavior meant to protect the organism from expending valuable energy. Of course, in a situation such as the squeezed rat scenario described by Seligman (1972), learned helplessness is maladaptive and lethal, so animals benefit only by striking a balance between energy- and life-saving behavior. Therefore, learned helplessness is not only a threat to an animal’s welfare, but also to survival.


Cooperative Care

            Training animals to opt into training sessions and treatments is commonly performed in the zoo setting and has more recently come into favor in companion animal veterinary medicine as well. The goal of cooperative care is to allow animals some control and predictability in routine care such as physical examinations, vaccinations, venipuncture, and grooming procedures, thus reducing stress, improving patient care, and preventing iatrogenic behavioral injury (Howell & Feyrecilde, 2018). Learned helplessness in the veterinary setting “can mask important findings on physical examination” (Howell & Feyrecilde, 2018, p. 2) and causing it is a direct violation of the veterinarian’s and veterinary technician’s oath.


Conclusion

            It is clear that though Weiss and Glazer (1975b) could not see a use for studying learned helplessness in animals, it has been useful in more contemporary research and applications in both humans and non-human animals (Howell & Feyrecilde, 2018; Conoscenti & Fanselow, 2019; Duncan et al., 2022). A study conducted on chimpanzees at the Johannesburg Zoo in 2009 yielded interesting results that suggested the animals may have experienced a form of learned helplessness (Duncan et al., 2022). The chimpanzees’ enclosure was renovated in 2004 and expanded from a few hundred square meters of barren space to 2500 square meters of enriched environment, encompassing the old habitat with major original structures intact (Duncan et al., 2022). All the chimpanzees had been living in the original enclosure for between 2-26 years prior to the renovation, and the researchers found the chimpanzees showed a preference for the areas where their previous enclosure existed and underutilized the remainder of the enclosure (Duncan et al., 2022). After a thorough review of potential explanations for this behavior, researchers could not rule out “spatial learned helplessness” and suggested further research in this area as the phenomenon has been demonstrated in other species as well (Duncan et al., 2022). Future research should seek to find what might predispose animals to learned helplessness and in what contexts; what factors of exposure are most important when avoiding learned helplessness; and effective humane methods for preventing and treating learned helplessness in companion animals.


 

References


Conoscenti, M. A. & Fanselow, M. S. (2019). Dissociation in effective treatment and behavioral     phenotype between stress-enhanced fear learning and learned helplessness. Frontiers in           Behavioral Neuroscience, 13(104). DOI: 10.3389/fnbeh.2019.00104

Duncan, L. M., D’Egidio Kotze, C., & Pillay, N. (2022). Long-term spatial restriction generates deferred limited space use in a zoo-housed chimpanzee group. Animals, 12(2207). DOI: https://             doi.org/10.3390/ani12172207

Ferrandiz, P. & Pardo, A. (1990). Immunization to learned helplessness in appetitive noncontingent          contexts. Animal Learning & Behavior, 18(3), 252-256.

Howell, A. & Feyrecilde, M. (2018). Cooperative veterinary care. John Wiley & Sons, Inc.

Moscarello, J. M. & Hartley, C. A. (2017). Agency and the calibration of motivated behavior. Trends in       Cognitive Sciences, 1704. http://dx.doi.org/10.1016/j.tics.2017.06.008

Overmier, J. B. & Seligman, M. E. P. (1967). Effects of inescapable shock upon subsequent escape and avoidance responding. Journal of Comparative and Physiological Psychology, 63(1), 28-33.

Seligman, M. E. P. (1972). Learned helplessness. Annu. Rev. Med., 23, 407–412. DOI:        http://doi.org/10.1146/annurev.me.23.020172.002203

Seligman, M. E., & Groves, D. P. (1970). Nontransient learned helplessness. Psychonomic Science, 19(3), 191–192. https://doi.org/10.3758/BF03335546

Weiss, J. M. & Glazer, H. I. (1975a). Effects of acute exposure to stressors on subsequent avoidance-         escape behavior. Psychosomatic Medicine, 37(6), 499-521.

Weiss, J. M. & Glazer, H. I. (1975b). Effects of chronic exposure to stressors on subsequent avoidance- escape behavior. Psychosomatic Medicine, 37(6), 522-534.

12 views0 comments

Recent Posts

See All

Commenti


bottom of page