In court, memory can be unreliable…because the brain behaves like a complex biological system.
Burlington, North Carolina, July 29, 1984. It is around 3:00 a.m. when a man breaks into the apartment of Jennifer Thompson-Cannino, twenty-two years old, and sexually assaults her. She survives, manages to escape, and calls for help.
In the hours and days that followed, she cooperated with the police, tried to reconstruct the attacker’s face, helped create a composite sketch, looked at photographs, and then identified Ronald Cotton. In court, that identification carried the weight of evidence.
Cotton was convicted. For years, the case seemed closed, based on what appeared to be solid legal truth. Then, in 1995, DNA testing proved that the man was not the culprit. After more than ten years in prison, Ronald Cotton was exonerated.
The Cotton-Thompson case has become one of the most studied examples in the psychology of testimony—not because there was a deliberate lie, but because something far more difficult to accept occurred: a sincere victim, convinced of her own memory, identified the wrong man.
Jennifer Thompson-Cannino had been the victim of a real crime. Ronald Cotton was innocent. In between—between the violence and the verdict—lay the fragile, powerful, and sometimes fallible realm of human memory.
It is in this space that the brain truly takes center stage before the judge. Not as a evocative image, not as a general mitigating factor, not as an explanation that fits every behavior, but as a complex biological system that perceives, selects, encodes, protects, integrates, erases, and reconstructs.
Justice requires facts; the mind, however, does not reproduce the past as an intact archive. Memory is not a neutral record of events, but a dynamic function, sensitive to stress, language, time, emotions, the relationship with the interrogator, and information acquired after the event.
Experimental research has shown this for decades. Elizabeth Loftus and John Palmer, in their famous 1974 study on eyewitness testimony, demonstrated how the wording of a question could influence the subsequent recollection of an event. Participants were shown footage of car accidents; they were then asked to estimate the speed of the cars.
The answer varied depending on the verb used in the question. It made a difference whether the question asked if the cars had “hit,” “collided,” or “crashed.” Language does not merely capture memory; under certain conditions, it shapes it.
In the legal context, this fact is decisive. A leading question, a detail introduced by the interviewer, confirmation provided too soon, or a photo lineup conducted without adequate precautions can affect memory more than common sense is willing to admit.
A witness does not simply open a drawer in their mind. They recall, organize, and put into words—often under pressure. And each step can alter the quality of the memory.
When a crime occurs under conditions of intense fear, the issue becomes even more delicate. The body under threat does not function to produce an orderly narrative; it functions to survive. The hypothalamic-pituitary-adrenal axis, the autonomic nervous system, catecholamines, and glucocorticoids all contribute to an alarm response that alters attention, perception, and the consolidation of the experience.
Theamygdala assigns salience to threatening stimuli; the hippocampus contributes to the spatial-temporal anchoring of the memory; the prefrontal cortex supports organization, evaluation, and regulation. When activation exceeds a certain threshold, some information may become strongly imprinted, while other details remain confused, peripheral, or intermittent.
This does not mean that trauma always erases memory, nor that every traumatic account must be accepted without verification. Rather, it means that a memory formed under extreme conditions cannot be evaluated using simplistic criteria.
A victim may recall a smell, a voice, a gesture, or a physical detail with precision, yet be unable to reconstruct the temporal sequence with the same clarity. They may remember the terror more than the duration, the threat more than the context, the weapon more than the face.
They may appear distant, cold, or contradictory—not because they are necessarily lying, but because, at the moment of the event, the nervous system prioritized defense over the orderly recording of facts.
In clinical practice, this is a well-known fact. In the courtroom, however, it is still often misunderstood. There is a tendency to imagine a credible victim according to an overly rigid script: they should report the incident immediately, remember everything, never change any details, react emotionally as expected, and appear neither too fragile nor too composed.
Yet the neurophysiological responses to a threat do not conform to an aesthetic of credibility. Freezing is neither consent nor moral inertia; it is an automatic defensive immobilization response. Dissociation is not a ruse; it can be an extreme form of protection when the experience exceeds one’s immediate capacity for integration.
The absence of a linear narrative does not in itself justify suspicion, just as suffering alone is not sufficient to constitute evidence.
Forensic psychological expertise serves precisely to avoid these two opposing errors: discrediting what does not appear orderly, or assuming that everything that is painful is true. The trial cannot turn vulnerability into a conviction of the defendant, but neither can it ignore the ways in which trauma alters perception, language, memory, and behavior.
Testimony is a mental and physical act: it arises from a situated perception, passes through an organism in a state of alarm, is recounted within a relationship, and finally enters the case file as words.
One of the most insidious issues concerns the relationship between subjective certainty and accuracy. A confident witness is convincing. A hesitant witness undermines credibility. However, research on eyewitness testimony has shown that confidence can be influenced by factors occurring after the event: explicit or implicit confirmations, repeated identification, the manner of the photo lineup, exposure to images, conversations, news reports, and the expectations of authorities.
A person may become more certain not because the memory is more accurate, but because it has been reinforced over time.
Wells and Olson, in their review of eyewitness testimony, drew attention to the role of misidentifications in wrongful convictions. The 2014 report by the National Academies of Sciences then recommended more rigorous procedures for identifications: blind or double-blind methods, standardized instructions, documentation of the degree of certainty expressed by the witness at the first identification, proper composition of lineups, and accurate recording of the entire process.
These are not mere administrative details. They are tools for safeguarding memory and, at the same time, justice.
Data from the Innocence Project confirm the severity of the problem: among DNA exonerations in the United States recorded between 1989 and 2020, 69% of cases involved a mistaken eyewitness identification. This does not cast suspicion on every testimony, but it does require that it be treated with caution. Witness error is not a marginal curiosity in criminology; it has led to actual convictions, years in prison, lives shattered, and victims forced to confront the possibility that they may have unwittingly contributed to an injustice.
Hypnosis, too, when discussed in relation to memory, requires a strict distinction. In the clinical setting, it can be used within a competent therapeutic framework to modulate anxiety, pain, attention, and certain psychophysiological responses.
In the forensic context, its use to “retrieve” memories must be approached with extreme caution. The literature on testimony refreshed through hypnosis has highlighted the risk of intrusions, increased subjective confidence, and the production of content that is not necessarily accurate.
This is not to discredit clinical hypnosis, but to remind us that treatment and evidence belong to different realms. A therapeutic setting is not an interrogation; clinical reprocessing is not the same as judicial certification.
The same applies to EMDR. The NICE guidelines on post-traumatic stress disorder list it, under specific clinical conditions, among the recommended interventions for adults diagnosed with PTSD or clinically significant symptoms. Its purpose is to facilitate the adaptive processing of traumatic memories and reduce the distress associated with the event. It is not intended to establish the truth in a legal proceeding, nor to produce evidentiary material. Confusing therapeutic efficacy with testimonial validity would be both a clinical and a legal error.
Neuroimmunomodulation helps us further understand that trauma is not merely a psychological phenomenon. Acute and chronic stress, sleep dysregulation, hypervigilance, attentional disturbances, vagal tone, energy metabolism, and resilience are all part of the same systemic architecture.
An organism exposed to a threat does not process the experience solely through thought: it processes it through the entire body.
For this reason, a traumatic testimony should not be heard as a simple account, but as the outcome of a complex interaction between memory, the nervous system, emotional state, language, and relational context.
In the contemporary world, the fragility of memory is compounded by additional factors of contamination. After an event, a witness may see photographs, read news reports, hear reconstructions, receive comments, discuss the event with others, and be exposed to videos, images, social media, and collective narratives. Each subsequent exposure can influence how the past is reorganized.
The procedural question, then, should not be limited to “What do you remember?”, but should include:
“When did you first recall it? “To whom did you recount it? What questions were you asked? Did you receive confirmation? Did you see images? Did you have access to subsequent information?”
Memory has a history, and that history must be understood.
None of this weakens justice; it strengthens it. Recognizing the limits of memory does not mean devaluing the victim, nor does it offer an undue advantage to the defendant. It means protecting both from oversimplification.
The victim has the right to be heard competently, without stereotypes and without suggestive pressure.
The defendant has the right not to be convicted on the basis of a subjective certainty that has not been adequately verified. The trial has a duty not to confuse the emotional power of a narrative with its probative value.
The brain, when presented before the judge, demands method. Method in the collection of initial statements, in photo lineups, in the formulation of questions, in the use of expert testimony, and in the distinction between clinical treatment and forensic evaluation.
It also demands restraint in the use of neuroscience. Merely citing the amygdala, hippocampus, or prefrontal cortex is not enough to explain an individual case. Science must enter the trial with restraint, not as a rhetorical flourish. It must clarify limits, not create bias.
Burlington, North Carolina, July 29, 1984. A young woman is assaulted and searches her memory for the face of the man who attacked her. She recognizes him. The court believes her. Ronald Cotton is convicted. Eleven years later, DNA evidence proves that certainty was wrong.
Since then, every time a witness utters the phrase “I remember”, the justice system should listen not only to the voice speaking, but also to the brain that produced it. Because memory can tell the truth.
But it can also be wrong without lying.
