False Memories and Inception
Our memories are so dear and essential to our lives that the idea of false memories or that our memories are vulnerable to outside influence seems more than a little unsettling. After all, our memories are critical for not only how we interact with others, but also generating and maintaining our image of self. The central theme of the movie “Inception” revolves around the ability to implant false memories in other people. While this makes for a captivating Hollywood story (and great movie), does this premise have any real scientific backing? Here, we’ll try to deconstruct how memories are represented in our brain and examine the possibility of implanting false memories in others.
Eyewitness testimonies in criminal cases provide a clue about the imperfect nature of our memory. Studies from the Innocence Project explain that in cases where people have been exonerated for a crime through DNA testing, faulty eyewitness testimony occurred more than 70% of the time that led to conviction.1 Assuming these were cases of honest misidentification, that is a staggering number to consider. With a process so fundamental to our lives, can it really be so fallible? Furthermore, can it also be manipulated by others?
The first attempt to describe memory appeared in Aristotle’s treatise “On the Soul” where he equated the brain to a blank slate (also called tabula rasa – if you really want to impress people at your next dinner party). In other words, “humans are born free of any knowledge and are merely the sum of their experiences.”2,3 Even 2,000 years ago, the importance of memory was not lost on individuals. Fast forward a millennia or so to the mid 1900’s where the academic community was intent on uncovering the biochemical trace of memories in the brain, or engram4. At its core, the engram is the idea that specific memories are represented in the brain by physical changes and thus can be traced. Why is this important for our purposes? If scientists could find physical evidence of a memory trace in the brain, imagine the implications of potentially erasing a memory (Eternal Sunshine of the Spotless Mind anyone?) or implanting a false memory. In a sense, the idea of the engram would make memory a physical, tractable entity in the brain rather than intangible ideas floating around in your head somewhere.
Karl Lashley was one of the most prominent figures in the early 1900’s that sought to uncover the engram. Lashley trained rats to find food rewards in a maze while lesioning different areas of their cortex. However, the locations of the lesions failed to correlate with the rats performance which led Lashley to conclude that memory was not localized to a specific area of the brain, but rather distributed throughout the brain in a theory known as equipotentiality. At one point, an exhausted and presumably discouraged Lashley proclaimed, “I sometimes feel, in reviewing the evidence on the localization of the memory trace, that the necessary conclusion is that learning just is not possible.”7
Despite this bold proclamation (which I’d imagine quite a number of people disagreed with), the learning and memory field chugged on and had a huge breakthrough in 1957 when Dr. Brenda Milner published a famous case study on patient H.M.8 In order to prevent his severe epilepsy, doctors bilaterally removed H.M.’s medial temporal lobe. While his epileptic seizures stopped, H.M. suffered from anterograde amnesia and temporally graded retrograde amnesia. The surgery effectively rendered H.M. unable to form new declarative memories9. That is, he could not remember or learn any new facts, places, or people. Finally, scientists could now localize an area of the brain (now known more specifically to be the hippocampus) required for the formation and storage of memories and knew that, despite Lashley’s slightly misguided claim, that human learning was in fact possible.
Now that we have narrowed down an area of the brain where memories are stored, we can consider the physiological processes of storing and recalling memories. The process by which newly learned information transitions to a stable long-term memory is known as consolidation. During the initial stages of consolidation, a memory is still vulnerable to outside events and can be disrupted with the application of protein synthesis inhibitors10. However, once consolidation is complete it is thought that the memory is permanent, impervious to manipulation or disruption (which could explain why H.M. retained some of his old memories)10. However, recent studies have challenged this long-standing belief with evidence of a process called reconsolidation. Reconsolidation is the idea that previously consolidated memories are, contrary to what I just told you, susceptible to disruption. While consolidated memories are insensitive to protein synthesis inhibitors, reconsolidation puts forth the idea that when a person recalls or remembers a memory it transiently opens up a window where the previously consolidated memory is now labile.
A clever study by Nader et al. (2000)11 used basic Pavlovian fear conditioning to test this phenomenon in mice. During training, mice learned to associate a tone (conditioned stimulus, CS) with a painful foot shock (unconditioned stimulus, US). Thus, once this association is learned mice will exhibit fear (scientists use mouse freezing as a metric for fear) when they hear the CS tone even in the absence of the shock. Notably, Nader and colleagues found that anisomycin (a protein synthesis inhibitor) application during a ‘reminder’ of the CS 24 hours after training produced amnesia in the mice. That is, the mice had forgotten the memory! When they applied anisomycin without giving a CS reminder they found the fear memory intact, suggesting that the CS reminder was a critical event that made the consolidated memory vulnerable. Now, if that’s not enough to get you jazzed up about science, the group took it a step further. They then repeated the same experiment, but now gave the reminder + anisomycin 14 days after mice had initially learned the memory (a significant period of time for mice). Even two weeks after mice had learned the association (when saline was injected they still exhibited high levels of freezing) anisomycin application during recall produced amnesia. So what does this mean?
Essentially, when you recall or remember a certain memory you enact an active process of reconsolidation that renders your memory labile again for a brief time. Thus, when you remember something such as an event, you are actively reconstructing the memory rather than just replaying it in your head. This could explain how our memories of certain events may become distorted over time. Example: on a personal level, this could be why certain concerts that I thought were just OK when I attended them have become a lot better in my memory over time. If every time I remember the show I think of how great it was, over many iterations my memory of the show is actively reconstructed and it’s no longer a “lie” I’m telling myself, but the actual representation of the memory in the brain reflects what I’ve been telling myself (so every time you think of this article, remember how great of a time you had reading it). Related: mental gymnastics.
So, if we can bend and shape our own memories, can a false memory be implanted by another? A paper from Dr. Susumu Tonegawa’s lab at MIT says yes (the paper is aptly titled Creating a False Memory in the Hippocampus12). The experiments used a lot of elegant mouse genetic tools, but here are the basics (also see fig 2). Mice were placed in a novel environment and allowed to explore while experimenters labeled the neurons that were representative of the context in the dentate gyrus, an area of the hippocampus (call this context A). Then, experimenters placed the mouse in a completely different environment (context B) and delivered a foot shock to the animals while simultaneously stimulating the neurons that represented context A with optogenetics. To quickly recap, experimenters artificially activated the pattern of neurons that once represented a “safe” environment with a foot shock even though the mice were in a different context.
Afterwards, mice were placed back in context A or another novel environment C. Although mice had never been physically shocked in context A, mice now exhibited significantly increased levels of freezing in A. As a result, experimenters were able to artificially engineer a memory such that a previously innocuous environment now scared the mouse just by manipulating neuronal activity. As a control, the mouse in context C was fine. These results lend credence to the idea of a memory engram. Through their genetic tools, experimenters activated specific ensembles of neurons during one behavior and were able to manipulate them so they were associated with a different behavior. In a sense, these mice were “incepted”. There seems to be a precise, physical trace of memories represented by patterns of neurons in the brain.
While many people simply misremember certain facts or events, I’ve tried to distinguish between someone with a bad memory from active processes in the brain that physically alter the long term physiological representation of the memory. One theory called the “Fuzzy Trace Theory” talks about false memories in the context of a mental battle between your actual memory and your capacity for reasoning13. Although there are many clinical cases of false memories, most are primarily dependent on similarities between the real and false memory rather than “true inception.”13 With evidence for processes such as reconsolidation, however, the physiological machinery for “true inception” to exist seems possible. We know that the hippocampus is a likely place to target, and that there are neural codes for specific memories in the brain. However, one of the most significant barriers in your endeavor to incept someone would be to create a non-invasive way to selectively manipulate someone’s brain activity. Until then, actual inception appears far off.
In sum, these examples have illustrated how labile and “flexible” our memories are – really just a euphemism for how our memories are not as good as we may think. While this might seem disappointing, consider how beneficial a flexible memory might be in order to forget misinformation. In fact, scientists are currently trying to take advantage of reconsolidation to treat ailments such as post-traumatic stress disorder14. Studies such as this to ‘hack’ the neural mechanisms of memory formation and elimination are ongoing and provide exciting possibilities for the future. Let us be glad that the scientific community did not heed Lashley’s claim. One thing we can conclude for certain…learning is indeed possible.
References 1. http://www.exonerate.org 2. http://www.human-memory.net/ 3. Aristotle. On the Soul, Book III, chapter 4. Prometheus Trust, 2003. 4. Lashley, Karl S. “In search of the engram.” (1950). 5. Lashley, Karl Spencer. “Brain mechanisms and intelligence: A quantitative study of injuries to the brain.” (1929). 6. Hunter, Walter S. “A consideration of Lashley’s theory of the equipotentiality of cerebral action.” The Journal of General Psychology 3.4 (1930): 455-468. 7. Lavond, David G., Jeansok J. Kim, and Richard F. Thompson. “Mammalian brain substrates of aversive classical conditioning.” Annual review of psychology 44.1 (1993): 317-342. 8. Scoville, William Beecher, and Brenda Milner. “Loss of recent memory after bilateral hippocampal lesions.” Journal of neurology, neurosurgery, and psychiatry 20.1 (1957): 11. 9. Squire, Larry R. “The legacy of patient HM for neuroscience.” Neuron 61.1 (2009): 6-9. 10. Davis, Hasker P., and Larry R. Squire. “Protein synthesis and memory: a review.” Psychological bulletin 96.3 (1984): 518.11. 11. Nader, Karim, Glenn E. Schafe, and Joseph E. Le Doux. “Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval.” Nature 406.6797 (2000): 722-726. 12. Ramirez, Steve, et al. “Creating a false memory in the hippocampus.” Science 341.6144 (2013): 387-391. 13. Reyna, Valerie F., and Charles J. Brainerd. “Fuzzy-trace theory: An interim synthesis.” Learning and Individual Differences 7.1 (1995): 1-75. 14. Debiec, J. A. C. E., and Joseph E. LeDoux. “Noradrenergic signaling in the amygdala contributes to the reconsolidation of fear memory: treatment implications for PTSD.” Annals of the New York Academy of Sciences 1071 (2006): 521-524.