Clues of memory lie dormant in every corner of life. We can see the silhouette of a loved one leaving in the corner of a wall bathed in sunset, smell the morning of parting with a relative in the biting wind, and taste the remnants of childhood in a Madeleine cake soaked in tea. Environmental clues can always activate the representation of certain memories in our brains. However, memory is not always filling our life experience; traumatic memories can also become inescapable nightmares for some people. Today, with a deeper understanding of the ‘three-inch black box’ in the skull, can we help people rewrite their traumatic memories of the past and regain the courage to face the future?
Rewriting memory used to be a plot in science fiction, but research on rodents (such as mice and rats) has proven that manipulating memory at the cellular level seems possible. In 2013, biologist Susumu Tonegawa led a team that successfully rewrote the memory of mice using optogenetics [1]. In their study, genetically edited mice were placed in Room A for free exploration. The genetic editing allowed the mice to express a light-sensitive protein in the “memory trace cells” of their hippocampus when exploring. On the second day of the experiment, the researchers placed the mice in Room B, applied a mild electric shock, and activated the memory cells encoding Room A with blue light. This operation replaced the mice’s memory, making them associate the electric shock with Room A. Therefore, in the final stage of the experiment, when the researchers returned the mice to Room A, they found that the mice showed a significant fear freezing response.
Although advanced technology has opened new paths for rewriting memory, the swapping of event associations in this study and the modification of complex human memories are still far apart, meaning that we are currently unable to apply this technology successfully to the human brain. The main challenge in rewriting memory is that the human brain has multiple expressions for the same event, each corresponding to complex neural representations. Therefore, directly stimulating the brain seems not to be an effective way to rewrite memory without being able to precisely locate memory representations. What else can we use then?
For a traumatic event (like a severe car accident), the brain may have multiple representations. One is the details related to the memory event, such as the weather and location of the car accident. These expressions are called episodic memory, which is generally stored in the hippocampus and then transferred or long-term transferred to the cerebral cortex; another is the physiological and psychological reactions triggered by the event. Environmental stimuli related to the car accident may cause the victim to repeatedly show stress responses or defensive behaviors, which are generally related to the amygdala; in addition, recalling the car accident may activate negative subjective feelings. Currently, most studies focus on changing these two expressions of episodic memory and stress responses to modify people’s traumatic memories.
After identifying the target of modification, how should we change people’s memory expression? Given the complex spatial network and wide distribution of memory representations, scientists first locate the time point when memory representations are more easily rewritten, i.e., when the neurons representing memory are activated. The time window when memory neurons are activated is generally divided into the consolidation period and the reconsolidation period of memory. The consolidation period of memory can be further divided into three stages: encoding, storage, and retrieval. During the encoding process, events are replayed through the hippocampus and stored therein. Then, this memory is processed by the prefrontal cortex and distributively stored in the neocortex. The consolidation period is relatively long and usually accompanied by repeated activation of neurons, which means that we can rewrite memory within a few minutes to a day after memory encoding.
Memory rewriting techniques during the consolidation stage usually target episodic memory regulated by the hippocampus and have changed the traumatic memories of human subjects to some extent. In addition, memories stored in our brains are highly dynamic and changeable. Each time after retrieval, the neurons representing memory will be activated for reconsolidation and then stored again. The reconsolidation period is another time window when memory is extremely vulnerable to rewriting. Each post-retrieval reprocessing may cause the original memory to change. Compared to the consolidation period, the reconsolidation period is shorter, so studies on manipulating memory during this period usually occur within hours after neuron activation.
Studies show that people usually remember events that trigger higher emotional arousal more firmly. [3] For example, when an event triggers our fear or sadness, our stress hormone levels will also rise, thereby enhancing the effect of memory consolidation. It was also found that even if the physiological arousal level of the subjects is increased after memory encoding, it can still improve the memory effect of the event. Therefore, when using memory rewriting techniques, drugs that can change the level of neurohormones are usually applied to enhance the effect of behavioral training. Currently approved hormone drugs for human subjects fall into two categories that can regulate memory consolidation: the first type of drug lowers the level of emotional arousal by blocking neurohormones, such as propranolol, which generally acts on β-adrenergic receptors in the amygdala, regulating memory consolidation occurring in the hippocampus; the other type of drug enhances memory consolidation by mimicking the action of neurohormones (such as adrenaline or glucocorticoids).
Most behavioral training combined with drugs targets the stress response triggered by traumatic events, but the effects of this technology are not stable. We know that immediate processing of the traumatic event may lead to secondary trauma and affect the victim for years. This may be because letting the victim recount the memory during the consolidation period of traumatic memory will increase their emotional arousal level, thereby enhancing the victim’s traumatic memory. [4, 5] So, can we prevent the consolidation of traumatic memories by administering blockers before psychological counseling? Studies have shown that administering propranolol to subjects immediately after a traumatic event can successfully lower the physiological arousal level when they recall the traumatic event but cannot effectively reduce their post-traumatic stress disorder level. [6]
The combination of mimicking neurohormones and behavioral therapy seems to be more effective. Administering mimicking neurohormones during the consolidation period can enhance the representation of event memory, strengthen the effect of exposure therapy, and ultimately effectively rewrite memory. The idea of exposure therapy comes from extinction learning: when people are repeatedly exposed to seemingly dangerous situations but do not suffer physical harm, they will gradually recognize that the situation is safe. Similarly, repeatedly recalling traumatic experiences may make victims realize that memories do not bring real physical harm. It should be noted that in this learning process, victims gradually replace their original traumatic memory with a new safe memory, rather than changing the original traumatic memory. Studies show that this replacement process may be related to the inhibition of the amygdala by the prefrontal cortex. [7] Since the expression of extinction learning is weak, the harm caused by traumatic memory may “resurface” at any time. To strengthen the consolidation of safe memory, some researchers will let victims use neurohormones while undergoing exposure therapy. Studies show that after adding drugs, the effect of exposure therapy can last for more than 6 weeks. Clinical results also found that administering glucocorticoids can effectively regulate subjects’ post-traumatic stress disorder response.
In addition to the above two drugs, there is another type of drug that combines the method of editing memory through extinction learning and uses the presentation of external cues. As mentioned earlier, memory consolidation is achieved by repeatedly activating the neuron cluster representing memory, which is usually related to the consolidation and integration of memory. Based on this, researchers developed targeted memory reactivation (TMR) therapy to regulate the stress response caused by traumatic events. We can apply an external stimulus to the subjects while they remember an event. Then, when the subjects are at rest or even sleeping, apply the same external stimulus to activate the same neural activity in their hippocampus that encodes the event. [8-10] TMR therapy can improve the stress response regulated by the amygdala. In a 2017 study, researchers presented an olfactory stimulus to subjects while applying a panic stimulus, causing the subjects to pair the two and produce a defensive response. Subsequently, researchers presented the olfactory stimulus again during the subjects’ slow-wave sleep. The experiment found that this operation led to the extinction of fear, and the subjects’ stress response was significantly reduced. [11]
Recent research shows that TMR therapy can not only weaken the stress response but also actively guide subjects to forget episodic memory. This study combines another memory editing technique called “active forgetting,” which means that by guiding, enhancing the subject’s motivation to suppress memory retrieval, thereby forming the extinction of a certain event memory. In the experiment, researchers let subjects learn the pairing of cues and targets representing different rules, one type of cue pairing target is subsequently reported, and the other type of cue pairing target needs to be suppressed for retrieval. After training, researchers will test whether the subjects have successfully learned these rules. Studies have found that people can suppress event memory retrieval through learned rules, and the number of suppressions is related to the level of forgetting. Repeated active forgetting can bring cumulative memory suppression effects. Neuroimaging results show that active forgetting is related to inhibitory regulation of the prefrontal lobe circuit, which may mean that active forgetting technology weakens the consolidation of traumatic memory by regulating the projection process from the hippocampus to the prefrontal lobe. [12]
The limitation of memory rewriting during the consolidation window is that this technology requires immediacy – the success window of neurohormone administration or behavioral marking techniques is within a few minutes to hours after the event, and the window of TMR therapy or active forgetting technology is longer, but it cannot be extended to several days later. Since victims usually seek help long after the traumatic event, such strict time requirements severely limit the application scenarios of the above therapies.
Research on the reconsolidation phase of memory has been conducted for nearly half a century. Nowadays, it is believed that reconsolidation generally has two functions: the enhancement and updating of memory. The former suggests that reactivating memory representations will improve memory effectiveness; the latter means that if other relevant stimuli exist in the environment at the time of memory reconsolidation, these new stimuli might be integrated into the original memory, thereby updating the memory. This has given scientists two ideas: if we adjust hormone levels during the reconsolidation period, we might be able to dissolve people’s memories of traumatic events; furthermore, if we present specific stimuli during this period, we might be able to rewrite the original traumatic memories into harmless new ones.
So, can we regulate stress responses during the reconsolidation period? Studies on rodents show that propranolol can effectively block the reconsolidation process and dissolve harmful memories. However, its effects are not stable in human subjects. Although administering propranolol before reactivation of the memory can reduce the level of post-traumatic stress disorder in people, it is unclear whether this effect is achieved by blocking the process of memory retrieval or by truly blocking the reconsolidation process of traumatic memory. If drugs are administered after memory reactivation, it does not effectively regulate stress symptoms.
More effective results come from sleep studies. During our sleep, rapid eye movement (REM) sleep accounts for about one-fifth of the total sleep time. Studies show that the brain usually replays events of the previous day during this period to reconsolidate memory. Sleep researcher Matthew Walker once heard from a doctor working at a veterans’ hospital that this doctor relieved the symptoms of nightmares in patients with post-traumatic stress disorder by administering a type of blood pressure medication. One of the side effects of this blood pressure medication is to lower the level of norepinephrine in the brain, a hormone closely related to our stress perception. Previous studies have shown that in normal subjects during REM, the level of norepinephrine in the brain is low, but this phenomenon is not observed in subjects with emotional disorders.
Walker speculated that in normal people, if the hormone level in the body is low during REM sleep when memory is reconsolidated, their emotional response might be alleviated. However, since patients with emotional disorders do not have this protective mechanism, the memory replay during sleep instead exacerbates the traumatic emotional response. Subsequent studies validated Walker’s hypothesis. Walker had one group of subjects watch a set of pictures at the beginning and end of the day, while another group watched a set of pictures before sleep and after waking up. Since the latter group reconsolidated emotional memory during REM sleep, their emotional response was alleviated when they viewed these pictures again. Lower activity levels were also observed in their amygdala. Further research found that due to the lower level of norepinephrine in the brain during REM sleep, reconsolidating memory during this “emotional safety period” might lead to a decrease in our emotional response to traumatic memories.
As mentioned earlier, extinction learning can affect stress responses. This training provides a new safe memory, a new competitor that can reduce the expression of traumatic memory, thereby reducing people’s stress responses. However, since this method does not change the original memory, there is a possibility that the stress response will reoccur. Studies show that extinction learning during the reconsolidation window of memory has better effects because we have the opportunity to truly rewrite traumatic memory in this window, rather than just providing a competitor for harmful memory. Studies on rats found that reactivating memory in the reconsolidation phase through reenactment conditions, followed by extinction training, can eliminate the stress response. By analyzing the brain activity of rats, researchers found that this type of extinction training triggered synaptic plasticity in the amygdala, rather than just inhibition of the amygdala by the prefrontal cortex. This may mean that this type of technology has successfully rewritten harmful memories. [13]
Subsequent studies show that this method is effective for human subjects. After reintegration, extinction training on the subjects can prevent the recurrence of defensive responses for at least one year. [14, 15] In subsequent neuroimaging studies, researchers further found that compared to ordinary extinction training, lower levels of prefrontal cortex activity were observed when training was conducted during the memory reconsolidation (retrieval) phase. [16] This is consistent with the conclusions of studies on rodents, which may also mean that the threatening memory represented by the amygdala has been edited, so inhibition by the prefrontal cortex is no longer necessary.
In 2000, neuroscientist Karim Nader from New York University proposed the role of the amygdala in the reconsolidation phase of memory, [17] which promoted the rise of memory research in this field. However, 20 years later, we have still not developed a clear and effective clinical treatment plan. Many significant studies cannot be replicated, and minor strategy changes may lead to vastly different experimental results. This situation on one hand shows the complexity of memory representations in the human brain, and on the other hand, proves that we need to understand the mechanism of memory more deeply before conducting practical exploration. Fortunately, many researchers have already taken steps in some directions, such as how to precisely locate the vulnerable window of memory, how the complexity of memory affects the difficulty of modification, or how to find neural markers that represent memory modification. On the other hand, we have not yet explored memory modification techniques that can be applied on a large scale, which may also reflect, to some extent, the limitations of behaviorism in the face of individual differences.
Today, we have developed many memory editing paradigms that are waiting to be applied. The development of brain science provides a brain mechanism for memory editing, and by deepening our understanding of this mechanism, we can further enhance the effectiveness of memory editing techniques in the future. With the rapid development of individual psychological and physical prediction models, as well as intracranial stimulation techniques, we have reason to believe that memory rewriting plans that can delve into each patient are on the horizon.
References
[1] The technique of optogenetics was proposed by Karl Deisseroth, which achieves the regulation of neuronal activity by expressing light-sensitive proteins in nerve cells, thereby allowing control with light stimuli of different frequencies.
[2] RAMIREZ S, LIU X, LIN P, et al. Creating a False Memory in The Hippocampus[J]. Science, 2013, 341(6144): 387-391.
[3] WOOD N E, ROSASCO M L, SURIS A M, et al. Pharmacological Blockade of Memory Reconsolidation in Posttraumatic Stress Disorder: Three Negative Psychophysiological Studies[J]. Psychiatry Research, 2015, 225(1-2): 31-39.
[4] MAYOU R A, EHLERS A, HOBBS M. Psychological Debriefing for Road Traffic Accident Victims: Three-Year Follow-Up of a Randomised Controlled Trial.[J] The British Journal of Psychiatry, 2000, 176(6): 589-593.
[5] SIJBRANDIJ M, OLFF M, REITSMA J B, et al. Emotional or Educational Debriefing After Psychological Trauma: Randomised Controlled Trial[J]. The British Journal of Psychiatry, 2006, 189(2): 150-155.
[6] HOGE E A, WORTHINGTON J J, NAGURNEY J T, et al. Effect of Acute Posttrauma Propranolol on PTST Outcome and Physiological Responses During Script Driven Imagery[J]. CNS Neuroscience & Therapeutics, 2012, 18(1): 21-27.
[7] DUNSMOOR J E, NIV Y, DAW N, et al. Rethinking Extinction[J]. Neuron, 2015, 88(1): 47-63.
[8] RASCH B, BORN J. Maintaining Memories by Reactivation[J]. Current Opinion in Neurobiology, 2007, 17(6): 698-703.
[9] TAMBINI A, BERNERS-LEE A, DAVACHI L. Brief Targeted Memory Reactivation During the Awake State Enhances Memory Stability and Benefits the Weakest Memories[J]. Scientific Reports, 2017, 7(1): 1-17.
[10] ALM K H, NGO C T, OLSON I R. Hippocampal Signatures of Awake Targeted Memory Reactivation[J]. Brain Structure and Function, 2019, 224(2): 713-726.
[11] HAUNER K K, HOWARD J D, ZELANO C, et al. Stimulus-Specific Enhancement of Fear Extinction During Slow-Wave Sleep[J]. Nature Neuroscience, 2013, 16(11): 1553-1555.
[12] HU X, BERGSTRÖM Z M, GAGNEPAIN P, et al. Suppressing Unwanted Memories Reduces Their Unintended Influences[J]. Current Directions in Psychological Science, 2017, 26(2): 197-206.
[13] MONFILS M H, COWANSAGE K K, KLANN E. Extinction-Reconsolidation Boundaries: Key to Persistent Attenuation of Fear Memories[J]. Science, 2009, 324(5929): 951-955.
[14] SCHILLER D, MONFILS M H, RAIO C M, et al. Preventing the Return of Fear in Humans Using Reconsolidation Update Mechanisms[J]. Nature, 2010, 463(7277): 49-53.
[15] https://www.the-scientist.com/careers/when-researchers-sound-the-alarm-onproblematic-papers-69086.[16] SCHILLER D, KANEN J W, LEDOUX J E, et al. Extinction During Reconsolidation of Threat Memory Diminishes Prefrontal Cortex Involvement[J]. Proceedings of the National Academy of Sciences, 2013, 110(50): 20040-20045.
[17] NADER K, SCHAFE G E, LE DOUX J E. Fear Memories Require Protein Synthesis in The Amygdala for Reconsolidation After Retrieval[J]. Nature, 2000, 406(6797): 722-726.
