Learned Helplessness

What it is

The work began in Richard Solomon's laboratory at the University of Pennsylvania, where Maier and Seligman were graduate students. According to the 2023 review, two senior students, Russell Leaf and J. Bruce Overmier, found by accident that dogs given earlier Pavlovian conditioning, a tone paired with footshock across 64 trials, later failed to learn a simple escape. The control dogs learned it easily. They only had to jump a small hurdle to end the shock.

Overmier and Seligman removed the tone and gave footshock alone in the first phase. The shocks could be neither escaped nor avoided, and the later escape deficit persisted. They suggested that "the source of the interference is a learned 'helplessness.'" The review credits them with the first use of the term.

Seligman and Maier then added the design that made the finding about control. The triadic design has three groups. In the escapable group, each shock ends when the animal presses a panel. In the inescapable group, each animal is yoked to a partner in the first group and receives physically identical shocks, which end only when its partner presses. A third group is not shocked. In the 1967 experiment there were three groups of eight dogs, and all were tested in a shuttle box a day later. The inescapable group was slower to escape, and in the authors' 1976 account six of its eight dogs failed completely. The review draws the conclusion that "prior shock per se did not produce interference, but only uncontrollable shock did."

The original theory held that organisms learn about the relation between their actions and outcomes. When shock is inescapable, action and the end of shock are independent, and the animal learns this and expects it to hold in future. Three mechanisms were proposed: a motivational one, in which the expectation that action brings relief is undermined; an associative one, in which learned non-contingency interferes with learning a contingency in a new setting; and a cognitive one, in which the animal comes to believe its responses are "of no avail."

The review lists what the theory handled badly. The escape deficit lasted only two to three days, which sits oddly with a lesson about uncontrollability. The authors had assumed uncontrollability was the active ingredient without testing whether control was. The theory explained failure to escape but not the wider changes that follow uncontrollable shock. And early methods could not show how brain activity depended on control.

In effect

Overmier and Seligman's 1967 paper reports three experiments on 88 dogs. In each, the dogs received inescapable shock in a first session and were given escape-avoidance training in a shuttle box about 24 hours later. All shocked groups escaped more slowly and failed more often than unshocked dogs, and different amounts of earlier shock did not differ from one another. A higher shock intensity during training did not remove the interference. Dogs paralysed with curare during the first session still showed it. In the third experiment, dogs tested 48, 72 or 144 hours after the shocks did not differ from dogs that had never been shocked, and the authors concluded that the effect dissipates rapidly.

The design reached people in 1975. In the account Maier and Seligman gave in 1976, college students heard a loud noise they could stop by pushing a button, heard the same noise with a button that did nothing, or heard no noise. All were then moved to a hand shuttle box in which moving a hand across the box stopped a noise. The students who could stop the first noise, and those who heard none, learned to use the box readily. Most of those who had heard noise they could not control failed to escape. The 1976 account also reports that students classed as externals on locus of control became helpless more easily than internals.

The revised account rests on rodent work, which the 2023 review describes. The standard paradigm uses mainly male rats in small wheel-turn boxes, with shock electrodes fixed to the tail. In the escapable group each shock ends when the rat turns the wheel with its front paws. In the yoked inescapable group wheel turning has no effect, and a third group is placed in the apparatus without shock. The authors first assumed uncontrollability was the active element and looked at serotonin neurons in the dorsal raphe nucleus, about 35 per cent of whose neurons in the rat are serotonergic. Inescapable stress, relative to escapable stress, activated those cells, raised extracellular serotonin and sensitised the neurons for about as long as the behavioural effects last. Lesions of the nucleus blocked the effects, and activating it without shock reproduced them.

The authors reasoned that the nucleus is too small, and lacks the sensory and motor inputs, to detect whether an action controls a shock. They proposed that the medial prefrontal cortex detects control and inhibits the nucleus. Cells in the prelimbic part of the cortex that project to the nucleus were activated after escapable shock, and silencing the cortex during escapable shock removed the differences in serotonin and behaviour without affecting the learning of the wheel turn. Activating the cortex during inescapable shock produced the protection seen after escapable shock. In the authors' words, control "actively leads to protection."

The same work reports immunisation. Escapable shock one week before social defeat prevented the defeat-induced rise in dorsal raphe serotonin and the behavioural deficits. Escapable shock in rats aged 35 days blocked the effects of inescapable shock in adulthood, 56 days later. Silencing the cortex at either time prevented the protection, as did blocking protein synthesis, NMDA receptors or MAPK. The review also offers a circuit explanation for the short time course that troubled the original theory: desensitisation of serotonin 1A autoreceptors, which it calls "a key molecular switch."

What it does not say

It does not say the 1967 experiments were wrong. The behaviour they reported has not been withdrawn by its authors. The 2023 review keeps the finding and replaces the account of why it happens.

It does not say the animals gave up for good. The dogs in Overmier and Seligman's third experiment looked normal after 48 hours, and the review describes the effect as lasting a few days.

It does not say the rat circuit has been shown in people. The authors call their translation to humans speculative and say it is uncertain which area of the human prefrontal cortex corresponds to the rat prelimbic region. They decline, in the review, to evaluate the human neural findings against their animal work. Until recently nearly all of the controllability work used male rats. In females, the review reports, control did not blunt the stressor in the same way: escapable shock produced the same outcomes as inescapable shock, and the authors write that "parallels remain unknown for humans and rats here."

It does not support clinical claims on this site. The 2023 review discusses implications for therapy and prevention, and those sections were not assessed for this page.

Popular retellings often stop at the original theory. Barry Schwartz's The Paradox of Choice, in its revised edition of 2016, says the animals "had learned from being exposed to inescapable shocks that nothing they did made a difference" (p. 106), which is the original account.

The 2016 paper by Maier and Seligman, "Learned helplessness at fifty: Insights from neuroscience," was confirmed from its record and not opened for this page. The account of the revision here comes from the 2023 review.

---