Are you the kind that remembers the core of a past event, but forgets the details? Well, research indicates that you might just be better at decision-making and adapting to the ever-changing, noisy environment. Most of us now acknowledge that it is as important to forget as it is to remember. And by forgetting, I do not mean wiping out unpleasant events (negative experiences propel better decision-making, we know that). It is storing the exquisite details or obsolete information that is a bother. Why? Picture this. Erin and Matsya are being taught to identify cubes. Each of them has a Rubik’s cube in hand and makes a mental note of the object. The Rubik’s cube is replaced with 3 objects — a dice, a sugar cube, and a multicolored ball, each of a different size. While Erin had kept in mind the Rubik’s cube color, pattern, shape, and size, Matsya only managed to recollect its shape. Simply by storing and applying the gist of the learning, Matsya could quickly predict the dice and sugar block as cubes (i.e generalize), whereas storing too many details impeded Erin’s ability to swiftly choose the cubes. In a different scenario, Matsya’s favorite ice cream shop in her neighborhood shifts to an adjacent locality. Ability of her brain to delete the old location and update the new one can avoid conflict between the old and new and ease her in finding the place. These two scenarios reflect the importance of having a right mix of memory retention and loss for optimal decision-making. Thus, the potential of memory doesn’t lie in accurate, long-term retention of information but rather in guiding sensible decisions and promoting a flexible/adaptable behavior.
The importance of memory transience has also been highlighted in machine learning (ML), an artificial intelligence approach, wherein machines are trained to learn from provided data and expected to self-improve their performance using the “learning”. Regularization, an ML process that is brain’s equivalent to ‘storing and applying the gist of the learning’, shows that the lesser the parameters used for modeling, higher is the model’s ability to correctly predict the outcomes of new data. On the other hand, overly accurate model systems that have too many fed-in parameters are lower in applicability as they cannot generalize over different data sets. Apart from regularization, computational models can also employ deletion of outdated data for more robust functioning. So, it looks like be it man or machine, remembering and forgetting are important.
But, what about the brain? What exactly is happening inside it when we are holding on to or letting go of memories? Can we influence what we retain or lose? Let’s take a quick look. The human brain is home to around 80-90 billion neurons — the smallest structural and functional electrically excitable units — that talk to each other using electrical and/or chemical signals. This “talking to each other” results in the formation of connections called “synapses”. Longer the talk between two neurons, stronger is their synapse (so much like human bonding, nay?). The birth, change, or death of these synapses is the basis for a lot of functions, one among them being storage and deletion of memories. Studies show that a memory persists principally because of excessive bonding between specific neurons that joined hands together to create the memory in the first place. Breaking or weakening of these bonds would aid in forgetting and/or learning. In reality, our brains are subject to regular remodeling from continuous neural activity and integration of new neurons. Moreover, environmental factors heavily influence our mnemonic abilities. For example, psychological stress affects an individual’s ability to store or retrieve memories, while activities like exercise are known to improve memory.
So, with memory’s neurobiological and computational perspectives in place, here’s the take home message: in a noisy, constantly changing world of today, optimal memory impermanence could be an investment in the choicest memory-guided planning for the future.
Reference:
Richards, B. A., & Frankland, P. W. (2017). The Persistence and Transience of Memory. Neuron, 94(6), 1071–1084.
About Saikata:
Saikata Sengupta is currently pursuing her Ph.D. from Department of Neurology at Friedrich Schiller University, Germany. You can follow her on Linkedin or Twitter.
Illustrator: Vinita Bharat, PhD of Fuzzy Synapse
Editors: Manoja Eswara, PhD and Paurvi Shinde, PhD
Manoja Eswara obtained her PhD from the University of Guelph, Canada and is currently pursuing her postdoctoral fellowship in Cancer Epigenetics at Lunenfeld Tanenbaum Research Institute, Toronto, Canada.
Paurvi Shinde did her PhD in Biomedical Sciences (Immunology) from the University of Connecticut Health and is currently a postDoc at Bloodworks Northwest in Seattle. Apart from science, she’s a trained classical dancer and loves outdoor and hikes.
The contents of Club SciWri are the copyright of PhD Career Support Group for STEM PhDs (A US Non-Profit 501(c)3, PhDCSG is an initiative of the alumni of the Indian Institute of Science, Bangalore. The primary aim of this group is to build a NETWORK among scientists, engineers and entrepreneurs).
This work by Club SciWri is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
