The Science of Memory, and how we’re Hoping to Rewire it

Oct 17, 2019 · 6 min read
Design by Eva Utke

Memory is neither perfect nor permanent: an array of studies have found that it is easily manipulated. The brain is not a “video recorder”, and even extraordinarily vivid memories have been shown to be inaccurate. Factors such as illness, trauma, substance use, or even nutrition, stress, and depression can damage memory. This imperfection has lead to a growth of innovation in neurotechnology seeking to address this issue: a prosthetic for the mind and a form of support for a degrading memory.

Creating technologies to support the formation and recall of memories requires an understanding of how these memories are constructed and stored. There are two types of memories: declarative and nondeclarative memories. Declarative, or explicit, memories can be consciously recalled and consist of facts and ideas a person encounters as well as their autobiographical memories concerning their experiences.

The creation of declarative memories is mainly managed by the hippocampus, which is associated with the arrangement of space within a memory. In coordination with other parts of the brain like the amygdala, which incites and is responsible for our emotional responses to events, the hippocampus assists in the formation of our working, short-term memories. These short-term memories are held in the prefrontal cortex. Eventually, these short-term memories may be consolidated to become accessible through the frontal lobe instead of the hippocampus, becoming long-term declarative memories.

Nondeclarative memories, which consist of skills and habits, are stored differently than declarative memories. They depend largely on the basal ganglia, a collection of structures near the base of the forebrain often referred to as the “habit center.”

As a result of consolidation, different memories may be accessed through different parts of the brain. Because of this, amnesia caused by brain damage rarely means the person has entirely lost all of their memories or their ability to create memories. It heavily depends on which areas of the brain are no longer accessible. Effective treatments for memory loss caused by brain damage are few and far between.

Elevvo, a non-invasive headpiece developed by BitBrain, is an emerging technology seeking to improve memory formation. Elevvo enhances cognitive skills — mainly memory, sustained attention and processing speed — through a series of intervention programs.

The precise therapy customized by Elevvo’s software has been proven to be effective in helping those with Major Depressive Disorder or ADHD improve the previously mentioned cognitive skills by as much as 30 percent. In the future, BitBrain hopes to apply this technology to help those with fibromyalgia, a musculoskeletal disorder that can cause issues with memory in addition to physical pain, as well as those with mild dementia.

The device itself is a portable, wireless EEG system taking under 5 minutes to set up. Using data it collects at the beginning of each session, a corresponding app modifies the program to accommodate the user. This treatment targets alpha brain waves — quick patterns of brain activity associated with alertness — and improves the brain’s ability to grow new neurons and rewire connections, called neuroplasticity. Following the completion of the entire program, users continue to check in every 3–6 months to determine if another round of intervention programs may help them more.

The exact mechanisms of memory formation that Elevvo enhances are far from understood. It is known, however, that this process depends on long-term potentiation (LTP) and long-term depression (LTD) of synapses. A synapse is a connection between two nerve cells. This connection can be strengthened or weakened by changing the number of receptive molecules responsible for communication between nerves. When this change persists, it is known as LTP or LTD and is a major component of the neuroplasticity targeted by Elevvo.

In addition to improving memory formation, novel neurotechnologies are looking to reverse the damage to memory caused by more severe physical trauma or dementia. Dementia is a prevalent example of memory deterioration, its most common form being Alzheimer’s. Alzheimer’s accounts for 60 to 70% of all dementia cases, and afflicts somewhere between 25–50% of the population over 85. The second most common form of dementia is vascular dementia, caused by the blockage of blood vessels in the brain. This is the same mechanism that causes a stroke, but it occurs “silently” in patients with vascular dementia — chronically impairing cognition, in contrast to the acute onset of a stroke. Other forms of dementia include Lewy Body dementia, Parkinson’s disease, and CTE, which — along with Alzheimer’s — are associated with the accumulation of certain proteins in the brain.

Most of the current treatments for memory loss caused by dementia consist of medications like cholinesterase inhibitors and memantine for Alzheimer’s or L-dopa for Parkinson’s disease. In response to damage and diseases that appear permanent, several companies are developing technology seeking to reverse the seemingly irreversible.

The Defense Advanced Research Projects Agency, or DARPA, supports a variety of research projects that may revolutionize how injured brains are treated. Its original project, Restoring Active Memory (RAM), was funded with the intent to help those who receive traumatic brain injuries during their military service.

Originally, RAM was meant to investigate the creation of a closed-loop, implantable system that would track how neurons create declarative memories and then, later, through targeted stimulation, improve the recall of the injured brain. Since an initial setback in research that resulted in the conclusion that the stimulation of the brain could not be purely based on targeted points, RAM has devolved into several offshoot sub-projects including RAM Replay and Nia Therapeutics.

RAM Replay is an offshoot of RAM that aims to support military training through the development of non-invasive neurotechnology that will help individuals develop and learn skills and habits more easily. Currently, the project is purely in the research phase. Researchers for RAM Replay are focusing on discovering a way to improve a person’s ability to acquire new skills and recall recent episodic memories by taking advantage of how the brain might “replay” memories while one sleeps. Studies have shown that sleep deprivation inhibits memory formation because deep sleep, also known as non-REM sleep, is vital to the formation of long-term memories by the hippocampus. RAM Replay hopes to capitalize on deep sleep with non-invasive technology to improve memory.

While RAM Replay explores the improvement of memory, Nia Therapeutics, founded by Michael Kahana and Dan Rizzuto, through funding from DARPA, has made headway in achieving RAM’s original goal of assisting those with injured brains recall newly formed memories with increased ease.

Nia Therapeutics is developing the Advanced Neurostimulation System (ANS), a closed-loop system powered by artificial intelligence that processes how a brain forms memories and then, based on this data, stimulates certain areas of the lateral temporal cortex to improve the user’s recall. The ANS, using information gained from previous studies identifying the specific brain regions — especially in the temporal lobe, which includes the hippocampus — and closed-loop systems that act as effective targets for therapy. Already, the ANS has been shown to improve patients’ short term memory of things like lists of items and sequential events by an average of 18%. With Nia Therapeutics’ recent acquisition of a chip developed by Cortera Neurotechnologies Inc., the capabilities of their Advanced Neurostimulation System is expected to increase even more. Once integrated into the ANS, Cortera’s chip will allow the ANS to track up to 64 different locations within the brain during the formation and recall of memories and then accordingly stimulate 4 different areas.

Our minds may be imperfect, but that does not mean that they cannot be improved. As scientists and engineers work together to better understand and support the brain, many of these afflictions may one day become a distant memory.

This article was co-written by Michael Xiong and Courtney-An Le and was edited by Jwalin Joshi

Michael Xiong is an undergraduate who studies Chemical Biology at UC Berkeley.

Courtney-An Le is an undergraduate student who studies Applied Mathematics at UC Berkeley.

Jwalin Joshi is an undergraduate at UC Berkeley, currently studying Applied Mathematics and Computer Science.


Writers, consultants, engineers, and designers working toward advancing neurotechnology for the benefit of humanity.


Written by

We write on psychology, ethics, neuroscience, and the newest in neural engineering. @UC Berkeley


Writers, consultants, engineers, and designers working toward advancing neurotechnology for the benefit of humanity.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade