Why Torture Doesn't Work: The Neuroscience of Interrogation*

In 2009, following the abuse of prisoners at its Guantanamo Bay detention camp, the US government made a significant decision. It moved the responsibility for 'enhanced interrogation techniques' from the CIA to a new government organization: the High-Value Detainee Interrogation Group (HIG). The move upset many CIA insiders; torture had been in their toolkits since the early days of the Cold War. The remarks of one official at a HIG-organized conference on torture in Washington DC can be summed up as: How could a new agency, created to both conduct and study torture, replace the decades of practice and perfection attained by the CIA? By adding a scientific component, responded the newly appointed head of the HIG.


Detainees were held at the United States' Guantanamo Bay detention camp in Cuba in 2002. Credit: Petty Officer 1st Class Shane T. McCoy/US Navy/Getty

This exchange highlights the theme of neuroscientist Shane O'Mara's Why Torture Doesn't Work. Rightly, O'Mara takes a moral stand against torture (forced retrieval of information from the memories of the unwilling). However, instead of simply providing utilitarian arguments, he argues that there is no evidence from psychology or neuroscience for many of the specious justifications of torture as an information-gathering tool. Providing an abundance of gruesome detail, O'Mara marshals vast, useful information about the effects of such practices on the brain and the body.

For instance, he explains why, physiologically, it is ludicrous to claim that stress, pain and fear will coerce a suspect to surrender critical information. The prolonged release of stress hormones such as cortisol damages the hippocampus — a brain structure crucial for encoding and retrieving memories — as well as the prefrontal cortex, which is implicated in decision-making and executive control processes. Such damage works in opposition to the goal of torture. Furthermore, chronic stress creates a negative feedback loop, causing enlargement and hyperresponsiveness of the amygdala, the brain structure that underlies emotional salience directs attention, enables learning and communicates with most of the brain.

Another striking example that O'Mara discusses is the effect on the brain of sleep deprivation. The practice was described in the 'Torture Memos' — legal memoranda drafted in 2002 by US Deputy Assistant Attorney General John Yoo, advising the CIA and President George W. Bush on the use of torture. Officially limited to a maximum of 180 hours and often combined with physical restraint, isolation, starvation and beatings, sleep deprivation has been used to coerce subjects into revealing information.

The memos further argue that sleep deprivation is harmless. O'Mara, however, discusses research suggesting that it erodes memory processes and general cognitive function by flooding the brain with glucocorticoid hormones. Even military scientists have produced literature that admits psychophysiological issues with sleep deprivation. In 1990, Paul Naitoh and his colleagues at the US Naval Health Research Center in San Diego, California, published evidence that the practice leads to an increase in circulating stress hormones and the development of psychomotor epileptic discharges (P. Naitoh et al. Occup. Med. 5, 209–237; 1990). They argued, too, that if combined with other stressors, such as food and water deprivation and waterboarding, sleep deprivation could negatively affect respiratory–and cardiovascular function.

Yet some officials and politicians continue to make announcements that run counter to such scientific evidence. Former Pennsylvania senator and Republican presidential hopeful Rick Santorum, for instance, commented in a 2011 interview that after being broken, people become cooperative. Most shocking maybe this year's revelation that a handful of officials in the American Psychological Association were complicit in torture by the United States after the September 2001 attacks on New York and the Pentagon, thus providing a veil of scientific legitimacy to the practice.

Torture also affects the torturer. The cognitive dissonance required to inflict suffering results in symptoms similar to those of post-traumatic stress disorder, O'Mara warns. He cites Joshua Phillips's None of Us Were Like This Before (Verso, 2010), which describes how many US veterans who had engaged in torture in Iraq experienced intense guilt or turned to substance abuse once back in the United States. Interviews with former interrogators in Northern Ireland, published by Ian Cobain in Cruel Britannia (Portobello, 2012), reveal that many believed what they had done was wrong but saw it as a desperate attempt to end the violence engulfing their society.

Conversation with a detainee may yield results comparable, and probably superior, to those obtained from torture.

Given that information obtained under torture is rarely reliable (because the victim will generally say anything to make the pain stop) O'Mara recommends an alternative: conversation. Having a conversation with a detainee may yield results comparable, and probably superior, to those obtained from torture. He cites three pieces of evidence.

First is a 1993 study by Stephen Moston and Terry Engelberg of police interrogations, which found that of more than 1,000 detainees, only 5% refused to talk (S. Moston and T. Engelberg Polic. Soc. 3, 223–237; 1993). Second, research by Robin Dunbar and his colleagues finds that 40% of what we reveal in conversation is 

related to the self, suggesting that refusing to self-disclose is very difficult (R. I. M. Dunbar et al. Hum. Nat. 8, 231–246; 1997). Third, a study by Diana Tamir and Jason Mitchell showed that people are willing to forgo money to talk to others about themselves. Indeed, the nucleus accumbens (part of the brain's reward circuitry) activates during such an opportunity, suggesting that people find disclosure intrinsically rewarding (D. I. Tamir and J. P. Mitchell Proc. Natl Acad. Sci. USA 109, 8038–8043; 2012). O'Mara does acknowledge that the difficulties of having such a conversation with a non-compliant person demand advanced social skills that are comparable to those of clinical psychologists and psychiatrists, who often deal with non-compliant patients. He suggests that alternative approaches, such as virtual reality and role-playing, may be useful for information gathering during interrogation.

Why, then, given its uselessness in eliciting valuable information, do people torture? It is a form of vengeance or punishment intended to discourage the victim from future transgressions and to communicate to others that harm will not be tolerated. In some cases, it occurs because the torturer believes that terrorists have mental illnesses. In science, however, punishment is not a viable response to someone with such an illness — just as torture is not a viable method for gathering information, as O'Mara repeatedly points out.

* Neuroscience: Tortured reasoning - Nature

Author information

Authors and Affiliations

Lasana T. Harris is a senior lecturer in experimental psychology at University College London and a guest lecturer in social and organizational psychology at Leiden University in the Netherlands. He studies the neuroscience of dehumanization and prejudice.

Lasana T. Harris

Corresponding author

Correspondence to Lasana T. Harris.