Its been a while. That’s because I got myself into trouble in the previous post by declaring a desire to sort out between two fundamentally different philosophies of how brains work. The more straightforward and more popular approach in brain research may be called the ‘outside-in’ framework.
The ‘outside-in’ framework considers the brain as an input-output device that receives sensory inputs and attempts to represent physical reality as faithfully as possible in neural activity called spikes. This sensory representation of the world is conveyed to the ‘thinking bits’ of the brain to be computed upon and the resulting output then commands the muscles or glands to generate an appropriate action (motor output in neuroscience jargon) in response to the sensory input.
The second approach, sometimes called the inside-out framework, puts emphasis on the brain as a constantly active, self-assembling network of networks whose ongoing neural activity is only modulated by its sensory inputs and, crucially, by its outputs as well. This framework focuses on the fact that the neural networks of the brain are constantly active even at rest or when asleep. This constant activity is predominantly governed by their own changing internal processes. Sensory inputs and copied output activity (feedback) only modify this on-going internal dynamics. An unknown, but significant proportion of these internal dynamics of the brain’s neural networks are governed by neuronal connectivity specified by our genomes that evolved over millennia. Crucially though, randomness and uniqueness are introduced into each brain by individual developmental conditions perhaps leading to unique neuronal networks in every adult.
Most practicing neuroscientists don’t necessarily think of their work in terms of the ‘outside-in’ or ‘inside-out’ frameworks, but the approach they pick tends to color everything, from the experiments they design to the conclusions they reach about data. It is useful to keep these frameworks in mind as we dig deeper into what is known about any particular neuroscience area.
I thought I would focus on how these two philosophies have been implicitly used to approach episodic memories, and our understanding of how the brain forms, stores, recalls and forgets them.
Little did I know, I was jumping into this deepest of rabbit holes in cognitive psychology and neuroscience. For the past several weeks I have been trying to reach the bottom of this seemingly endless hole. But, sorry to say, I am not there yet. Here is my first attempt, with no guarantees that I have even got the shape of the field right! Much of the following discussion is grounded in the ‘outside-in’ framework predominant in cognitive psychology and neuroscience, except the speculative discussion later about how episodic memory may have evolved which is also not devoid of ‘outside-in’ thinking but at least considers evolution as a factor.
In my opinion, the most impressive feat of memory most humans exhibit, is the ability to recall vivid details of events that happened to them. I’m less drawn to superhuman feats of memory a few people can perform, like recalling pi (3.1415..) to 100,000 digits. These people are like olympic level athletes whose abilities are beyond reach for most of us. I’m more interested in vivid recall about long ago personal events that most humans possess. A memory of a past event one has been directly involved in, is known as episodic memory. In some ways, this is even more remarkable than the mental gymnastics of recalling pi to 100,000 digits or more, since each episode happens only once but can leave lasting traces for the duration of a lifetime.
Endel Tulving, one of the most influential cognitive psychologists of the twentieth century, coined the term episodic memory in 1972 to differentiate these memories from semantic memories. To clarify the difference using an example: your knowledge that Paris is the capital of France is a semantic memory, but your memory of your trip to Paris last summer recalling the hotel you stayed in, the restaurants, cafes and museums you visited, the food you ate, the Parisians you interacted with, are part of your episodic memory.
A subset of episodic memories are autobiographical; a set of individual, unique memories and facts like birth details, mother tongue, favorite food etc. which we recall and narrate when asked to describe ourselves. Autobiographical memories can last a lifetime and can be remarkably detailed accounts of the episode, including subjective percepts like colors, smells and even feelings recalled several decades after the actual event.
Here’s a bunch of randomly selected people on the street recalling their earliest memories:
(Tip: turn on close captioning and speed up video to 1.25x).
If you listen to the stories above, you begin to notice certain common patterns. These memories are selective, nobody says they remember every day of their childhood! The memories are from decades ago, yet very vivid; they are all narrated from the first person perspective which means narrating what they experienced with their own senses, feelings and perception rather than as a third party observer. They have high emotional content and are very sensory in nature, based on sights, sounds, smells etc.
The first narrator (0:08 in the video above) talks about picturing running around his grand mother’s house as a kid and is transported there; he is re-smelling the old furniture, his grandmother’s suffocating perfume; re-experiencing the mood of that day running around the house with his siblings and cousins. The second narrator (0:57 in video) explaining how she misspelt a word, relives the shame and discomfort she felt and then consoles herself with the remark that everyone makes mistakes. At 3:20 in the video, there is a white haired, older lady recalling vivid details of her mother’s stove accident when she herself was in a crib, a memory that is easily from several decades earlier! The last guy (8:20 in the video) recalls being three and lying on his dad’s chest in a hammock, relives the safety and comfort of that moment.
It appears that each one of them is traveling back in time and experiencing the episode with some or all their senses, perceptions and emotions. Every memory is narrated in the first person perspective; smells, feelings like nausea, safety, comfort; emotions like shame, disgust, happiness dominate these narratives as if in some advanced VR (virtual reality) game that not only has visuals but also includes other sensations, feelings and emotions as well.
Writing thirty years after coining the term, Tulving tried to boil down the unique character and requirements of episodic memories into a short list: a sense of subjective time: a past that begins at one’s own birth (in reality, at the mid-point or second-half of the first decade of life in humans) up to the present moment in one’s own life; being aware of the past and recognizing that this awareness is different from being aware in the present moment (he called this autonoetic awareness) and thirdly, a self: if an episodic memory is mental time travel, then the traveler is the self within each of us.
Tulving surmised there was no evidence other species possessed these three capabilities and concluded no other species is capable of episodic memory. Since this conclusion depends on people narrating their own subjective experiences -which is impossible for other species to do- it remains unfalsifiable. Current experimental evidence strongly suggests other species do possess at least some form of episodic memory and has opened the door to experimental investigations of the neural basis and evolutionary origins of episodic memory1.
The vividness with which people recall past events is one of the most fascinating aspects of episodic memories. This vividness has been suggested as being an important input for planning future actions2. The first-person perspective allows us to imagine what might happen to us if we undertake a certain action. For tree-climbing primates like us, one might imagine this scenario:
“the last time I climbed this tree, I saw a huge, scary beehive (recalling fear within an imaginary situation: classic first-person perspective) on one of the higher branches. Should I climb it? What if the bees attack me and I fall to my death? Maybe I should search for another tree to climb.”
The fear resulting from the first-person perspective seems crucial to avoid accidental harm or death due to a planned action.
In general, one can think of episodic memory and imagination as mental time travel to the past and to the future, respectively. Was this the reason why our lineage evolved episodic memory? Not as a faithful record of past events, but as a tool to help predict future actions?
Furthermore, this insight that episodic memory may be important in planning future actions suggests it evolved from brain systems responsible for navigating the physical environment the animal is in. This is called spatial navigation in the jargon and includes capabilities that allow animals to move about safely in a physical environment filled with hidden and unknown dangers along with valuable resources located at various places, thus requiring it to map the space and plan the next action. This capability seems crucial for survival.
These ancient brain mechanisms may have been co-opted by evolution to allow us to navigate other, more abstract spaces. Consider social relationships for example. We can construct a mental map of all our cousins and their families in our minds, as well as making a mental map of our neighborhood. The abstract map of your cousins allows you to navigate your extended family relationships, while the map of the neighborhood allows you to navigate your neighborhood. These maps called ‘cognitive maps’ were first proposed by Edward Tolman in the 1940s when he was studying rats navigating mazes. Subsequently, tantalizing suggestions of how brain structures like the hippocampus implement these maps (the closest we are to nailing down the connection between the brain and the mind) have been uncovered through decades of painstaking neurophysiological and neuropsychological experiments. More on this in the next post.
Despite this postulated benefit for survival and procreation, episodic memories can be quite unreliable as factual records of the past. People swear to what they saw, heard or otherwise experienced several years ago, when serving as eyewitnesses in legal proceedings. The Innocence Project has documented 375 exonerations so far, based on DNA evidence of people wrongfully convicted of crimes they didn’t commit. 259 (69%) of these convictions were due to mis-identification by eyewitnesses. Clearly these strong memories formed in the heat of the moment aren’t very reliable as facts.
There is a body of psychological literature on false memories and how other people, including psychology researchers, can implant false memories into subjects. They range from relatively innocuous ones like adverse reactions to certain food items leading people to avoid those; to life-changing false memories in children about long-term abuse by guardians, ironically implanted by psychotherapists asking suggestive questions3. It is believed that false eyewitness testimonies are many times the result of being influenced by biased or premature media reports as well as suggestive law enforcement questioning. Several eyewitness testimonies began as very tentative identifications that became much more confident by the time of the trial suggesting a change in the original memory of the incident.
Interestingly, our autobiographical memories (a subset of episodic) contribute to our self-image; something which has been acknowledged since the time of William James, the father of American psychology. Most of us might be generally satisfied with our self-image but a negative self-image can be the reason why some people are depressed and can even loathe themselves. Autobiographical memories are a large part of who we think we are. There are other types of memories (e.g. procedural memories like how to ride a bicycle) which are much more reliable than episodic memories, yet those unconscious memories don’t influence our autobiographical self as much as episodic memory seems to. Why are episodic memories, and particularly autobiographical ones, malleable in this way? Does this mean that our self-image is partly or mostly fictional? Dependent on perhaps increasingly false memories of events that get distorted a little bit each time they are recalled?
People’s ability to form and retain autobiographical memories spans a spectrum; from recalling almost every day of their past to having very few or no autobiographical memories. A tiny fraction of humans have autobiographical memory capacities far beyond most other people. These individuals with ‘highly superior autobiographic memories (HSAM)’ were studied systematically by the prominent memory researcher James McGaugh and colleagues at the University of California at Irvine.
Some HSAM individuals are able to recall the day of the week, personal and public events that occurred, newspaper headlines, even the weather for many or most of the days since they were around 10 to 11 years old! See a few of them demonstrate their abilities in the video below, though the tentative conclusions reached by the researchers about the brain structures of these individuals are quite preliminary. In recent times, the sample size of HSAM individuals in this study is considered too small for reliable conclusions regarding brain structure.
Jill Price, the original subject of McGaugh’s studies, became somewhat of a celebrity with appearances on early morning news shows and CBS’s 60 Minutes, with audiences in the millions.
The UCI group was subsequently contacted by a large number of people claiming to possess memory like Jill’s. Using a battery of interviews that extensively probed their memories of personal and public events (a more systematic but similar process to the TV interviews shown in the clip ), they identified 30-40 verified HSAM individuals. They eventually tested twenty HSAM individuals and an equal number of matched controls for a range of cognitive tasks (see this paper4 to learn about the tests) in an effort to identify the cognitive abilities that contribute to HSAM.
The UCI group were trained psychologists and in doing these tests were implicitly using the ‘outside-in’ framework. These tests assume a rationally designed memory device which -prior work in psychology suggested- may ‘reasonably’ thought to possess certain capabilities like: paying attention to the episode, ignoring distractions, working memory, recall of visual images recorded earlier, emotional engagement etc. They then tested each of these capabilities, summarized in the table above, to identify the superior ability of HSAM individuals over control subjects to provide a ‘logical’ explanation for HSAM.
Unfortunately they didn’t discover any individual capability to be greatly superior in HSAM individuals over those of the control subjects. Vastly superior autobiographical memory capability seemed to be very specific and did not extend to even standard memory tests conducted in laboratories or clinics! For example, HSAM individuals didn’t score any better than matched controls on standard memory tests like recalling words from arbitrary word-pairs they were exposed to just prior to their test. To quote from the paper:
“The results suggest that superior cognitive functioning is an unlikely basis of HSAM, as only modest advantages were found in only a few tests.”
Given their incredibly accurate recall of personal and public events over several decades, this failure to capture their superior autobiographical memory in the standard tests suggests - in my opinion- these assessments are not engaging the relevant brain mechanisms that underpin HSAM.
At the other end of the autobiographical memory spectrum are individuals with almost no autobiographical memories. This condition is referred to as ‘severely deficient autobiographical memory’ (SDAM). SDAM individuals lead normal high-functioning lives, don’t show any brain abnormalities but report an inability to relive or vividly recall past events of their own life, seemingly the opposite of HSAM individuals. This lack of vividness contradicts the earlier speculation that vivid first person experiences are necessary or helpful in planning future actions (the beehive on tree generating fear above) since these SDAM individuals seem to do fine without it. What then is the biological function of the first person perspective in episodic memories?
When tested for various cognitive abilities, SDAM subjects showed no differences from control subjects except on a visual memory task that involves recalling a complex figure from memory, 30 minutes after being allowed to study it. The recalled complex figure drawn by a control subject and by three SDAM subjects (AA, BB and CC) is shown below.
The three SDAM subjects in the study above are clearly deficient in their memory of the complicated figure. This lack of visual recall has been suggested to be the reason for the lack of vividness in their autobiographical memories, as well as the reason for their lack of, or poor, autobiographical memories.
Although research into HSAM and SDAM is just beginning, these conditions pose a puzzle regarding autobiographical memories. Both these types of individuals seem to get along fine in real life. Most of us fall in between these extremes but don’t seem to suffer either due to lack of HSAM or, do better than individuals with SDAM. A naive interpretation might be : autobiographical memory is not really crucial for human survival, the way regular, non-autobiographical episodic memory and spatial navigation are.
On the other hand, vivid autobiographical memories are where many of us experience the richness of our inner lives. People who develop amnesia due to brain injury or degenerative disease certainly experience a much poorer quality of life. How do we explain this conundrum? Our brain seems to endow us with these mysterious autobiographical memories which seem to be not strictly necessary for our biological lives?
Allen, T. A., & Fortin, N. J. (2013). The evolution of episodic memory. Proceedings of the National Academy of Sciences, 110(supplement_2), 10379-10386.
Buckner, R. L., & Carroll, D. C. (2007). Self-projection and the brain. Trends in cognitive sciences, 11(2), 49-57.
Lynn, S. J., Lock, T., Loftus, E. F., Krackow, E., & Lilienfeld, S. O. (2003). The remembrance of things past: Problematic memory recovery techniques in psychotherapy.
LePort, A. K., Stark, S. M., McGaugh, J. L., & Stark, C. E. (2017). A cognitive assessment of highly superior autobiographical memory. Memory, 25(2), 276-288.