The Neurological Basis for the Theory of Recovered Memory

The Neurological Basis for the Theory of Recovered Memory

Neil D. Brick MA Ed. – June 2003

Copyright 2003 – All rights reserved. No reproduction of any material without written permission from the author.

author contact e-mail address

This paper will present research showing the biological basis for the theory of recovered memory. I will define recovered memory as the phenomenon of partially or fully losing part or a specific aspect of a memory, and then later recovering part or all of the memory into conscious awareness. This paper will include data from the works of van der Kolk and Fisler, Knopp and Benson and Bremner. Supporting data will include PTSD studies on Vietnam veterans and survivors of childhood trauma, subjective reports of memory, measurements of stress-responsive neurohormones, animal research on neurohormones, neuroimaging and MRI brain studies. Different theories of memory and amnesia will also be presented.

van der Kolk and Fisler’s research shows that traumatic memories are retrieved, at least at first, in the form of mental imprints that are dissociated. These imprints are of the affective and sensory elements of the traumatic experience. Clients have reported the slow emergence of a personal narrative that can be considered explicit (conscious) memory. The level of emotional significance of a memory correlates directly with the memory’s veracity. Studies of subjective reports of memory show that memories of highly significant events are unusually accurate and stable over time. There are a variety of memory systems which usually operate outside of conscious awareness. These systems operate with some independence from the other memory systems. While people appear to easily assimilate expected and known experiences, aspects of traumatic experiences appear to get stuck in the mind, unaltered by time passing or experiences that may follow. The imprints of traumatic experiences appear to be qualitatively different from those of nontraumatic events. Explicit memories of personal facts or events are affected by lesions of the front lobe and hippocampus. These parts of the brain are also involved in PTSD neurobiology. Traumatic memories may be coded differently than ordinary event memories, possibly because of alterations in attentional focusing or the fact that extreme emotional arousal interferes with the memory functions of the hippocampus. (van der Kolk & Fisler 1995)

Traumas can interfere with several memory functions. van der Kolk divided these functional disturbances into four sets, traumatic amnesia, global memory impairment, dissociative processes and traumatic memories’ sensorimotor organization. Traumatic amnesia involves the loss of remembering traumatic experiences. The younger the subject and the longer the traumatic event is, the greater the chance of significant amnesia. Global memory impairment makes it difficult for these subjects to construct an accurate account of their present and past history. Dissociation refers to memories being stored as fragments and not as unitary wholes. Not being able to integrate traumatic memories seems to be the main element which leads to PTSD. In the sensorimotor organization of traumatic memories, sensations are fragmented into different sensory components. (van der Kolk & Fisler, 1995)

van der Kolk and Fisler’s study of 46 adults supports Piaget’s notion that when memories can’t be integrated linguistically or semantically, they are organized in a more primitive manner as somatic sensations or visual images. In a collaborative neuroimaging study with the authors, it was found that when subjects had flashbacks in the laboratory, there was increased activity in the right hemisphere in areas connected to the processing of emotional experiences and in the right visual association cortex. Broca’s areas in the left hemisphere showed significantly decreased activity. This helps back up the theory that traumatic memories are made up of emotional and sensory states with minimal verbal representation. van der Kolk and Fisler’s hypothesis is that under extreme stress, the memory categorization system based in the hippocampus fails, allowing for these memories to be kept as emotional and sensory states. Excessive arousal at the moment of trauma interferes with the clear memory processing of the event, leaving unaltered memory traces. When these memory traces are remembered and put into a personal narrative, they are subject to being condensed, contaminated and embellished upon. When memory traces are recalled, increased activity in the amygdala might cause the personal assignment of accuracy and individual significance. (van der Kolk & Fisler, 1995)

In van der Kolk’s work on the psychobiology of PTSD, he states that trauma responses are bimodal. The response to trauma is hypermnesia, over reaction to stimuli and reexperiencing the trauma, which exists with numbing, avoidance and amnesia. When compensating for chronic hyperarousal, behaviorally subjects with PTSD may shut down and avoid stimuli similar to the trauma. Psychobiologically, they may emotionally numb both to trauma and everyday experience. PTSD subjects may go directly from stimuli to response without an adequate appraisal of the situation, due to their overgeneralization of incoming stimuli, creating flight or fight reactions. (van der Kolk, 1994)

Abnormal physiological responses in PTSD have been shown in two ways. One is due to reminders of the trauma. The second is due to intense, neutral stimuli, such as loud sounds. Individuals with PTSD show several autonomic responses to these stimuli, like blood pressure, skin conductance and heart rate. These highly elevated responses show the timelessness and intensity of how traumatic memories may affect one’s present experience. Lang proposed that emotional memories are stored as associative networks, where these networks are activated when a person gets confronted by a sufficient number of elements that make up these networks. Kolb proposed that excessive stimulation of the CNS during trauma could cause permanent neuronal changes. These changes would have a detrimental effect on stimulus discrimination, habituation and learning. (van der Kolk, 1994)

Abnormal acoustic startle response (ASR) has been a main feature in trauma response for more than 50 years. ASR in PTSD subjects response is mediated by such excitatory amino acids as glutamate and aspartate. ASR is modulated by several neurotransmitters and second messengers at supraspinal and spinal levels. Abnormalities in habituation are found in ASR with PTSD subjects. This failure of habituation for PTSD to loud sounds suggests problems with evaluating sensory input. The fact that PTSD subjects are unable to properly integrate trauma memories is shown physiologically by their misinterpretation of nonthreatening stimuli. (van der Kolk, 1994)

PTSD develops following intense stressors. Intense stress causes the release of stress-responsive neurohormones, like cortisol, norepinephrine, epinephrine, etc. Constant exposure to stress changes an organism’s adaptiveness and how it deals with its daily environment. Studies have shown neuroendocrine abnormalities in PTSD subjects. These studies have shown chronically increased activity of the sympathetic nervous system activity in PTSD. The neurochemicals measured in these studies include catecholamines, corticosteriods, serotonin and endogenous opioids. Through self-reports of emotional responses, it is suggested that endogenous opioids are responsible for the numbing of emotions in response to trauma. Putnam’s work shows large neuroendocrine changes in sexually abused girls when they are compared to normals. (van der Kolk, 1994)

Trauma victims do not respond to stress the way normals do. Pressure situations may cause a feeling of retraumatization. High states of arousal may promote the retrieval of trauma memories and associated phenomena such as sensory information or behaviors connected to prior trauma. Therefore, traumatic memories may be considered state dependent. Under stress, people secrete endogenous stress hormones that affect memory consolidation strength. Through studies on animal models, it is assumed that the large secretion of neurohormones during a traumatic event in part causes long-term potentiation (LTP) and the over-consolidation of traumatic memories. This LTP may cause an organism to remember a trauma whenever aroused. Neuroepinepehrine may be the major hormone causing LTP. Endorphins and oxytocin may actually cause inhibition of the consolidation of memories. Reliving the traumatic event may cause stress hormones to strength the memory trace causing a positive feedback loop. (van der Kolk, 1994)

The limbic system is believed to be critically involved in memory storage and retrieval as well as giving emotional significance to sensory inputs. Research in brain imaging studies suggests that trauma patients may have limbic system abnormalities. One part of the limbic system, the amygdala, may assign free-floating feelings to input which are then elaborated upon by the neocortex and imbued with personal meaning. It may also integrate internal representations of the external world in memory image form associating emotional experiences with these memories. The septo-hippocampal system is thought to record memory in temporal and spatial dimensions, and plays an important role in storing and categorizing incoming stimuli in memory. Hippocampal damage is connected to over responsiveness to external stimuli. When stress interferes with the hippocampus’ mediation of memory, it is possible that some of the memory is kept by a system that records emotional experience, but there is no symbolic placement of it in time or space. In animals, high stimulation of the amygdala interferes with hippocampal processing. Strong affect may disallow proper evaluating and categorizing of an experience. (van der Kolk, 1994)

Hebb distinguished between short-term and long-term memory. He postulated that any memory that stayed in short-term storage for a long enough time would be consolidated into a long-term memory. Later research showed this to be false. Research has shown that direct injections of cortisol or epinephrine help the storage of recent experiences. This is also true for stimulation of the amygdala. This proves that excitement enhances memory by the stimulation of hormones that affect the amygdala. Excessive or prolonged stress (with prolonged cortisol) may hurt memory storage. Patients with amygdalar damage are no more likely to remember emotionally charged words than nonemotionally charged ones. Baddely and Hitch developed the term working memory to show that temporary storage is more than a station on the way to long-term memory, it is the way we store memory when memory is worked with or examined. The hippocampus is important for explicit memory. The hippocampus is also important for memory consolidation. The hippocampus receives input from different parts of the cortex and sends it’s output out to different parts of the brain also. The input comes from secondary and terciary sensory areas that have processed the information a lot already. Hippocampal damage may also cause memory loss and problems with memory storage. (Kalat, 2001) This data backs up the idea that the hormones caused by excitement may enhance memory. An example of this is being able to remember sexually charged words. However, excessive stress and hippocampal damage may damage memory processing. This is somewhat similar to what van der Kolk described.

Kolb and Whishaw describe two groups of theories about amnesia. The defects in sequential processing theory states that to get a memory to become permanent, there needs to be a structural change in the brain. Amnesia is assumed to occur when the consolidation process is disrupted and either the memory trace doesn’t produce a structural change and gets lost or there is a structural change and access to the memory trace is lost. The multiple memory systems theories state that amnesia occurs due to a problem in one of the memory systems. These theories assume that the impairment of a psychological system may cause deficits in some kinds of memories but not others. Kolb and Whishaw claim that none of these models can explain all of the aspects of amnesia. (Kolb & Whishaw, 1995) van der Kolk’s theory of traumatic amnesia appears to be a combination of both of the above theories. In the defects in sequential processing theory, amnesia occurs when the consolidation process is disrupted, and access to the memory trace is lost. Amygdalar disruption of hippocampal processing may cause this consolidation disruption in terms of the proper categorization of a memory. The multiple memory systems theories state that impairment may be caused in one system but not another. van der Kolk’s work theorizes and attempts to prove that there is a difference between traumatic and nontraumatic memory storage and retrieval.

The body’s need to respond in danger situations can be strong. There is a tremendous physiological and neurochemical cost to this type of response, due to the depletion of hormones and neurotransmitters. With adequate recovery time, the body can return to its own homeostasis. When there is inadequate recovery time between stressful situations, alterations may occur to the neurophysiological parts of one’s stress-response system. Some of these alterations may be irreversible. One’s body’s memory and learning systems may be altered affecting implicit and explicit memory. This may cause maladaptive or pathological responses. This damage may cause memory loss, learning deficits and other maladaptive symptoms. Children’s neurological and physiological systems are very vulnerable to the negative impact of trauma. Highly resilient people may have a better chance of experiencing a trauma without developing PTSD. But if a trauma is strong enough, no person is immune to the consequences of developing PTSD. Uncontrollable stress may have a similar impact biologically. It may be possible to look at an individual’s pre and post trauma neurochemistry and tell if they have experienced trauma, but it would not be possible to say what kind of trauma. Animal studies show us that learned helplessness can develop from repeated exposure to inescapable trauma. In humans, physical paralysis has been shown to be a main feature connected to a traumatic event. This paralysis has been shown to be connected to hypermnesia, amnesia and dissociation. Traumatic events may be unavailable to recall or may be recalled only in pieces. (Knopp & Benson, 1996)

Cathecholamine activity may be altered due to trauma. The intrusive symptoms of PTSD are thought to be connected to the dysregulation of the cathecholamine system. Several systems are regulated during the cathecholamine systems reaction to stress. These include, the locus ceruleus/norepinephrine system, which affects the activities of the cerebral cortex, hypothalamus, hippocampus and amygdala, the dopamine system, which causes and increase in activity in the prefrontal cortex, the SNS, which causes physical arousal. (Knopp & Benson, 1996)

Amnesia involves losses in explicit memory. It is shown by one’s inability to remember personal memories or discuss them verbally, or it may be shown by one’s inability to fully retain in conscious awareness temporarily retrieved memories. Amnesia is often considered to be a dissociative condition, such as dissociative amnesia. An individual may only remember parts of the event, or certain categories about the event (like feelings). Amnesia caused by deficits while encoding information may not be reversible, because the information was not encoded. Gaps in autobiographical memory are normal to PTSD sufferers, as are problems with nonstressful short-term memory tasks. The successful coding of memories entails alert focused awareness when the input is presented. Memory consolidation is most successful when the experience can be elaborated on in conscious thought. A lack of conscious awareness may hurt these processes. Extreme elevations of norepinephrine released in trauma situations are related to the strong implicit hypermnesia memories and to the explicit deficits of memory of amnesia. Medium to high levels of norepinephrine cause the amygdala to promote LTP in the hippocampus, which may result in vivid memories. Very high levels of norepineephrine and heavy stimulation of the amygdala connected to extreme, prolonged or repeated stress appear to interfere with hippocampal functioning. This interference may hurt cognitive assessment and the encoding of the input. (Knopp & Benson, 1996)

van der Kolk writes about the neural connections between the hippocampus and cerebral cortex. They are the main pathways of communication between subcortical and cortical areas. These connections can be assumed to represent a link between unconscious and conscious brain areas. Changes in the hippocampus’ functioning during uncontrollable stress may hurt and limit the consolidation of the input into the explicit memory system. Some mental representations of the input may remain in cortical emotional memory, which may cause phobias and free-floating anxiety. This explains how trauma sufferers may have amnesia for specific events, but not the emotions connected to them. Excessive levels of opioids released in the brain during trauma and the numbing response connected to them may also be a major factor for the impairment of memory. According to van der Kolk, in animal studies, memory is damaged when a situation can no longer be helped by the animal’s activity. Panic and freeze responses may be seen as defensive ways to allow an organism to not consciously experience overwhelming stress or to not remember an occurrence of overwhelming stress. High opioid levels may influence the encoding of input in two ways. The first is by changing the initial experience of one’s perceptions by decreasing the perception of pain and changing other senses. The second is by changing one’s interpretation of the entire event, creating an unreal context for the event, like physical or emotional detachment. These events are characteristic of dissociative responses. These influences may cause memories that are unrelated to or dissociated from the normal methods of explicit memory retrieval. (Knopp & Benson, 1996)

Bremner’s work states that changes in the neurotransmitter GABA in response to uncontrolled stress have been connected to losses in memory and learning. This may be due to GABA’s interrelationship with neuroepinephrine and opioids. Chronic dopamine dysregulation may be a consequence of trauma. Dopamine’s relationship to working memory in the prefrontal cortex may also show its connection to problems of encoding and short term memory. High levels of cortisol (as mentioned previously by Kalat) may cause memory deficits because of its neurotoxic effects on the hippocampus. In animal studies, high levels of cortisol have been shown to cause hippocampal damage. In humans, MRI studies have shown reduced hippocampal volumes in combat veterans with PTSD, adults with posttraumatic symptoms and survivors of repeated childhood sexual or physical abuse. Hippocampal damage may cause short-term memory retention deficits. van der Kolk writes that PTSD’s essence is memory disturbances. Janet wrote that certain occurrences may leave intense memories in a person. With dissociative trauma survivors, trauma may also interfere with implicit memory, where periods of avoidance may be interrupted by intrusive emotional occurrences with no story to guide them. (Knopp & Benson, 1996)

Bremner cites several studies showing a connection between hippocampal volume and stress related disorders. The hippocampus is sensitive to stress and plays an important role in memory and learning. The hippocampus also has a role in recording the emotions of a stressful event. Researchers have measured hippocampal volume with MRIs. In one study, hippocampal volume was found to be reduced by 8% on the right side in Vietnam combat veterans with PTSD. Later a 12% reduction was found on the left side. After this, Bremner measured left hippocampal volume in patients with PTSD related to early childhood sexual and physical abuse and also found a 12% reduction. Other studies have replicated these findings. Abuse and PTSD are related to a broad range of memory disturbances, and PTSD sufferers may be more susceptible to memory problems than normals. A difficult issue for Bremner is whether those presumably abused accurately recall their information. ( Bremner, 2002)

Changes in memory function by cortisol and norepinephrine may show us a mechanism for the delayed recall of child abuse. Cortisol over hours acts to weaken the laying down of memory traces and neuroepinephrine may strengthen these traces. Exaggerated cortisol release in a stressful situation in PTSD may cause the inhibition of the retrieval of memories. The exaggerated release of norepinephrine in animal studies could be connected to humans. This would show how memory recall can be facilitated. This may also explain how traumatic memories of childhood abuse could suddenly break into consciousness and would back up the claims of PTSD patients. Frontal lobe abnormalities causing dysfunction may also underlie the delayed recall of abuse memories. Studies show an increased propensity for memory distortions in women with self-reported abuse and PTSD. Prefrontal cortical dysfunction in PTSD caused by abuse could be a reason for these phenomena. This could explain an increase in the capacity for source amnesia effects. Failure of activation or decreased blood flow in the medial prefrontal cortex may cause increased fearfulness inappropriate for the situation. This inability to regulate emotions could cause PTSD sufferers to avoid reminders to protect themselves. This could lead to amnesia. It has been shown that PTSD symptoms increase after the delayed recall of childhood abuse. ( Bremner, 2002)

In conclusion, there is a great deal of strong scientific data to show a neurological basis for the theory of recovered memory. Van der Kolk shows how memories may be dissociated into different parts, how traumatic memories may be remembered differently and how neurochemicals may affect the amygdala and the hippocampus, causing memory problems. His research on neuroimaging appears back up his theory that traumatic memories are made up of emotional and sensory states with minimal verbal representation. This would support the theory that traumatic memories may be remembered in fragments. van der Kolk’s research also shows us the possibility of stress causing damaging to the memory systems that help categorize memories. Kalat’s book also backs up the theory that neurochemicals may enhance or damage memory processing. Kolb and Whishaw’s book discusses different theories of memory. van der Kolk’s work appears to consider more than one theory of memory, because his theory discusses both how the memory consolidation process is disrupted and access to the memory trace is lost, and that there is a difference between traumatic and nontraumatic memory storage and retrieval. Knopp and Benson’s book appears to back up van der Kolk’s work. It discusses how traumatic events may be unavailable to recall or may be recalled only in pieces. Knopp and Benson’s book also states some of the neurochemical processes that may cause impairment of the memory process. Their book discusses the theory that memory consolidation is most successful when the experience can be elaborated on in conscious thought. Bremner’s work shows us how certain neurochemical levels may cause memory impairment. He also shows a connection between reduced hippocampal volume and certain kinds of trauma. Bremner also theorizes about the connections between brain neurochemical levels and the delayed recall of child abuse. Hopefully, future scientific breakthroughs will continue to help researchers collect data on the neurological basis for the theory of recovered memory, so that sufferers of severe trauma will be able to fully receive the treatment and support they need and the lingering scientific question about the existence of recovered memories will finally be put to rest.


Bremner, J. D. (2002). Does stress damage the brain? New York: W.W. Norton and Company

Kalat, J.W. (2001). Biological psychology (7th ed.). Belmont, CA: Wadsworth Publishing.

Knopp, F. H. & Benson, A. R. (1996) A primer on the complexities of traumatic memory childhood sexual abuse; a psychobiological approach. Brandon, VT : Safer Society Press

Kolb, B., & Whishaw, I. (1995). Fundamentals of human neuropsychology (4th ed.). New York: W.H. Freeman.

Schacter, D. L. & Scarry, E. Eds. (2000) Memory, brain and belief. Cambridge, MA: Harvard University Press

van der Kolk, B. A. (1994). The body keeps the score: Memory and the evolving psychobiology of post traumatic stress. Retrieved August 17, 1999 from

van der Kolk, B. A. & Fisler, R. (1995) Dissociation and the fragmentary nature of traumatic memories: Overview and exploratory study. Retrieved August 23, 1999 from