The Possibilities of BrainMesh

An incredible team at Harvard developed BrainMesh — a really thin mesh that can be rolled up and then injected onto the surface of the brain. The mesh has properties similar to neural tissue and can be completely accepted by neural tissue in mice. Once in the brain, it can wire up brain regions, while focusing on individual neurons — with the ability to read and write.

Importantly, this has only been tested in mice thus far, and we do not yet know if the mesh can be life-long, although the results are on the way. ¹ Nonetheless, with the high-precision of BrainMesh, it holds immense promise and potential for the questions we may be able to answer in the human brain. In this article, we’d like to pose three possible gaps in research that the BrainMesh could shed light on — neurolinguistics, the neuroscience of death and the Bayesian brain hypothesis.

Exploring Neurolinguistics

Language is integral in allowing us to communicate our experiences in and with the world around us, but it also has the power to shape our perceptions as well. An important question facing the world of Neuroscience is the extent to which language affects how we think and perceive the world, and the neural mechanisms underlying this process. One curiosity is bilingualism. If one speaks two languages that represent the world differently, then how do both languages combine to create their perception of the world? What about languages that conceptualize time differently?² Does the bilingual individual toggle between both languages or do both concepts exist abstractly in their mind and then they retrieve whichever concept they wish? Crucially, how does this phenomena get represented in the brain? To what extent do the neural networks for the different languages overlap? What may be the neural mechanisms underlying how different languages combine to shape our perception?

Newer technologies like BrainMesh can help answer these questions in a way that current technologies cannot. Because of its highly portable nature, information can be continually gathered while the subjects engage in tasks that allow them to engage in the world with a freedom that current neuroimaging devices cannot offer. We can better track how the brains of monolinguals across two separate languages are activated when navigating their way around a city and then compare that to a bilingual who speaks both languages compared to someone laying in an fMRI machine or hooked up to an EEG machine who is navigating through a simulation. We can also use this device to track language acquisition and development in children being raised in bilingual households and in older children and adults learning a second language and observe how their brain changes throughout that process. We may then assess how their perception of the world across specific variables such as perception of time, differ from monolinguals or across early age bilinguals and bilinguals who learned a second language at a later stage in life. It is thought that there is a general language network, within which each language has its own variation and through BrainMesh we can track those variations as subjects switch from one language to another.³ This technology holds endless possibilities for observing the neuronal behavior of bilingualism.

The Neuroscience of Death

We have many anecdotal stories of near death experiences (NDE) that give us a glimpse into what the process of death might feel like. We also have an idea of what the brain goes through in traumatic death, but with newer neuroimaging technologies we may be able to understand exactly how the brain processes death and how it is physiologically affected by such an event. While the stories of NDE offer clues, they don’t tell us everything. The biggest issue is that there are many situations where an individual may experience an NDE but they actually experienced a syncope, also known as fainting, or fear that triggered NDE-like experiences that were grouped in with true life threatening experiences.⁴ Granted, the experiences were similar across both true NDEs and harmless situations but with much data on NDEs including experiences of individuals who did not actually encounter or were close to death, we would still like to exclusively look at individuals who can show us how death is experienced in the brain.

While death due to trauma is often an acute event there is the question of whether or not natural death is an event or a process. With BrainMesh, we can study the brains of healthy seniors and track the days or weeks prior to passing to observe any neural correlates indicative of approaching death. Postmortem analysis of the brain may also be conducted for a considerable amount of time after death to track the degeneration of neuronal cells, specifically how long it takes the cells to die, their activities and the chemicals they release, and the extent to which there is still communication among neurons between brain regions and to the rest of the body. We can also assess which brain regions shut down first, the order in which the brain regions and cells die, and try to better understand the psychological and subjective experience of death interpreted through the brain’s behavior. Because of the intimate and often unexpected nature of death, traditional neuroimaging techniques are inadequate to capture the experience but with BrainMesh, the device is attached to the surface of the brain at all times so when an event occurs, it will be captured and the information stored for later analysis.

Euphoria during Death and Near-death experiences

A state of euphoria has been documented in patients during their final moments, especially in the absence of pain-killers. There is evidence of a serotonin rush in rats near the moment of death.⁵ It also seems that pain declines during the dying process, even weeks before the subject actually dies, although some experience the same amount of pain just a few hours before their passing.⁶ Euphoria has also been seen to occur in subjects with near-death experiences. Scientist Jill Bolte Taylor explains so in her TED Talk after she had a stroke in her brain’s left-hemisphere.

How could this be explained? Maybe the flood of endorphins triggered by the “dying process” is a last-minute survival mechanism to try to wake up the body. Or maybe it is the depletion of the energy we had left, for the same purpose. Some recognize the power of placebo, suggestion, religious beliefs and meditation techniques in overcoming the pain during the dying process. There is also evidence that suggests that injury in the brain’s right-hemisphere is partly responsible for the feeling of a “higher-power”, which could increase the chances of euphoria or tranquility (except if you worry about the consequence of your terrenial sins).⁷ It could even be a way for our bodies to give us a small treat before we depart, although how this could have evolved to be is not intuitive.

As a whole, we still do not have a clear understanding of the subjective experience of dying, a process that all of us will eventually go through. BrainMesh could be extremely helpful as a tool to allow researchers to record activity changes in many brain regions at once, which is important if we want to have an integral idea of how the brain experiences the dying process. BrainMesh could be implanted, for example, in healthy young mice and then used to track their brain’s activity up to their death, including recordings of their dying process and even after their passing. Thinking further in time, an updated version of BrainMesh could also help keep track of non-electrical activity in the brain, such as the state of glial cells and the activity and complexity of neurotransmitters, both in the healthy and dying subjects. Still, to have a more clear and conclusive understanding of the neural mechanisms underlying the process of dying, there needs to be higher funding for death-related investigations and palliative care research.

The Developing Bayesian Brain

The Bayesian brain hypothesis posits that our brains are essentially prediction machines.It argues that the brain is constantly inferring the causes of sensory input based on our internal model of the world. Now, how does this relate to perception? With the brain as a prediction machine, its hypothesized that perception is the result of the top-down inferences of causes of sensory input, rather than being purely driven by bottom-up sensory signals. ⁸ Importantly, this internal model can be updated and optimized with prediction errors, which arise when the predicted sensory data and actual sensory data don’t match.

However, there still exists a large gap between the Bayesian brain hypothesis and our understanding of the neural mechanisms that could potentially underlie it. With the Bayesian brain hypothesis also comes the question of how we come to develop our internal predictive processing models, especially during the crucial years of infancy and childhood.

What if the potential of BrainMesh could be harnessed to shed light on the neural mechanisms underlying the developing Bayesian brain? The level of specificity that BrainMesh allows for, the reading of individual neurons, could allow us to understand how the Bayesian brain develops.

Firstly, Series and Seitz (2013) discusses the possibility that structural expectations could be reflected in the tuning of neurons to the features that we are expecting.⁹ With this, we might also expect that neurons representing the most expected features would be more sharply tuned or in larger numbers. Some studies support this by showing that the enhancement of neuronal activity in the lateral intraparietal area of monkeys reflected the most likely response. With this hypothesis, what if BrainMesh could be used to map out how the tuning of individual neurons changes over the course of development? And how does this reflect the development of an internal predictive model?

Another intriguing idea is that spontaneous activity could be reflecting our brain taking samples of the prior distribution. Series and Seitz (2013) explains that with sensory inputs, the posterior distribution is a result of the sensory input and the prior distribution, i.e. our current beliefs about the environment. Without any sensory inputs, spontaneous activity will correspond to the prior distribution. This is argued to be more advantageous as it shortens our reaction times since the network is already at a state that corresponds to the most likely inputs. Hence, another question with how our predictive models develop is if BrainMesh could potentially allow us to understand how sampling of the prior distribution changes over development? This could also be linked to periods of increased learning and mapping the learning to changes at the neuronal level.

Importantly, research will have to begin in the context of mouse brains before being able to observe what happens in the human brain. And it is on the team’s radar to begin understanding how infant mice brains may react to BrainMesh, and how it may integrate with the developing brain.

Closing Question

While it may be a few years away, BrainMesh has immense potential to address important questions that we’ve been pondering as a society with an extremely high precision that we’ve never had before. With the ability to read and write from individual neurons in the human brain, what are questions you’d like answered?

References

1. Dailymail.com, C. (2017). Radical ‘brain mesh’ that could make the Matrix a reality. Retrieved November 06, 2020, from https://www.dailymail.co.uk/sciencetech/article-4668722/Radical-brain-mesh-make-Matrix-reality.html
2. Does Language Shape Thought?: Mandarin and English Speakers’ Conceptions of Time. (2001). Cognitive Psychology, 43(1), 1–22.
3. Wong, B., Yen, B., & O’Brian, B. (2015). Neurolinguistics: Structure, Function, and Connectivity in the Bilingual Brain. Retrieved 1 November 2020, from https://sci-hub.do/10.1155/2016/7069274
4. Nelson K. (2015). Near-death experiences — Neuroscience perspectives on near-death experiences. Missouri medicine, 112(2), 92–98.
5. Elevation of brain serotonin during dying. (2011). Neuroscience Letters, 498(1), 20–21.
6. Symptom prevalence in the last week of life. (1997). Journal of Pain and Symptom Management, 14(6), 328–331.
7. Johnstone, B., Bodling, A., Cohen, D., Christ, S. E., & Wegrzyn, A. (2012). Right Parietal Lobe-Related “Selflessness” as the Neuropsychological Basis of Spiritual Transcendence. In International Journal for the Psychology of Religion (Vol. 22, Issue 4, pp. 267–284). https://doi.org/10.1080/10508619.2012.657524
8. Seth, A. K. (2015). The Cybernetic Bayesian Brain — From Interoceptive Inference to Sensorimotor Contingencies. In T. Metzinger & J. M. Windt (Eds). Open MIND: 35(T). Frankfurt am Main: MIND Group. doi: 10.15502/9783958570108
9. Seriès, P., & Seitz, A. R. (2013). Learning what to expect (in visual perception). Frontiers in Human Neuroscience, 7. doi:10.3389/fnhum.2013.00668