A conversation with Dr Samuel Barnes by Jean Rintoul
Jean Rintoul is a first year PhD student in Nir Grossman's lab. She investigates novel strategies to achieve better focal resolution in non-invasive neuromodulation. Prior to joining the UK DRI Jean worked in industry on projects such as the Emotiv EEG BCI system, an open source electrical impedance tomography system, and a bioimpedance spectroscopy solution for neurogenic bladder disorder.
Dr Samuel Barnes is a UK DRI fellow investigating why the aged brain is vulnerable to neurodegeneration in order to identify strategies that may alleviate this susceptibility.
Jean: Sam, thank you for taking time for this interview. To start with, tell us a little bit about your career background and research interest.
Sam: Yes of course. I graduated from Oxford University in 2006, and then I took up an MRC capacity building PhD studentship at King’s College London in the Institute of Psychiatry working with Dr. Gerald Finnerty. In that position I was investigating the mechanisms of neural plasticity that drive synaptic connection loss in the cortex. It was during this time that I became very interested in the neuroplasticity mechanisms that prevent runaway activity levels. I was using electrophysiology at the time which can be a painstaking way to gather information about a neural circuit as complex as the cortex, and I realised that using imaging approaches was a much more tractable solution to the problem. I therefore decided to move to UCL and took up a Post Doc position in Professor Tara Keck’s group investigating how homeostatic control of neural firing rate is regulated in the adult visual cortex using in vivo imaging. I had a great time there and really enjoyed the topic, people around me and the skills I developed. Then in 2015, I was lucky enough to win the Edmond and Lily Safra Fellowship position and moved to Imperial College London in the Department of Brain Sciences. This was an independent fellowship which supports early career researchers as they transition from being a post doc to independence. It’s a fantastic position, and a couple of people in the department now hold a similar post. I think it’s a great way to develop your career. I was lucky to have a very supportive group of people around me to help me do that including Professor Thomas Knopfel and Professor Paul Matthews. I published a few papers from that post and then in April 2018 I was awarded the UK DRI fellowship position and became a lecturer in the department of Brain Sciences. My group is interested in the role of plasticity mechanisms in both health and disease. We study homeostatic control in both the aged brain and pre-clinical models that capture features of Alzheimer’s pathology. We want to know why the aged brain is vulnerable to AD and how dysregulated neural circuit plasticity and spiking activity might drive synapse loss. Our goal is to try and develop new interventions that can slow or prevent neurodegeneration. To study this question, we use a combination of in vivo voltage and calcium imaging, bioelectronic stimulation, electrophysiology and automated behavioural experiments.
Jean: Very interesting. Can you describe a little bit about homeostatic plasticity processes, and what got you interested in them?
Sam: Yes of course. Homeostatic plasticity mechanisms basically control neuronal firing rate. Homeostatic control is the cornerstone of all biological systems. It’s absolutely paramount to an organ like the brain to prevent excessive neural spiking levels, dysregulated calcium, changes in wiring and dysregulated activity patterns across the brain. It’s a fundamental feature of the brain. So, the question itself, is what are the key mechanisms? Well, if you just take a single neuron there are actually multiple mechanisms. The reason we think there are multiple mechanisms, is that controlling the firing rate is so important that the system has to provide redundancy in order to prevent a catastrophic failure of one system leading to runaway excitation or periods of prolonged quiescence. One of the key mechanisms I’m particularly interested in is the control of synaptic strength. Synaptic strength can be dialled up or down in relation to activity of the neuron, in order to prevent conditions of highly elevated or very low activity. Failures in the ability to regulate synaptic strength may contribute to early stages of some diseases. Some of the other mechanisms involve regulation of intrinsic neural excitability. This can occur due to changes in ion channel expression, so that output spiking can be modulated or the balance between excitatory synaptic input and inhibitory synaptic input. This is a really important mechanism that we think is disrupted in the early stages of AD. Non-neuronal cells such as microglia and astrocytes also play an important role in surveying the activity of the neural network and releasing certain cytokines and other modulating molecules to coordinate plasticity. We think the interplay between these multiple homeostatic mechanisms and cell-types is important for understanding multiple disorders, but particularly for Alzheimer's Disease and cognitive changes associated with brain ageing.
Jean: It reminds me of over damped and under damped control systems like the PID or proportional integral differential controllers used to guide robot arms, which are all about how do you get a stable system that doesn’t run out of control or just damp down to nothing.
Sam: Yeah, absolutely. We see the overshoot and undershoot controller models happening on a network level. So many of us in the field of homeostasis, think about homeostatic control from that kind of engineering perspective to try and test new hypotheses in brain plasticity.
Jean: This homeostatic pattern that is destabilizing in Alzheimer's patients, is this something that you can differentiate from schizophrenia, or how does it appear different from other disease pathologies?
Sam: Well, that's a very interesting question. So obviously, the basic pathologies of those two disorders are very different, though there are some similarities in the pattern of regional brain dysfunction. First of all, we need to test if homeostatic control is failing in the early stages of Alzheimer's disease. My group is trying to better understand how homeostatic control in is perturbed by the aging process and by increased amyloid. From this point, it would be interesting to see different disorders have a common dysregulation of activity or homeostatic control mechanisms.
Jean: Right. So how do you recognize these abnormal electrical patterns?
Sam: We’ve been spending a lot of time thinking about this question. We characterize the activity patterns and microcircuit interactions in both wild type animals of different ages and preclinical models. We do a very careful characterization of the sensory response properties, their resting state activity, their activity at different anaesthetic levels, and how stable activity patterns are over long periods of time. We then perturb the system, using a variety of different approaches such as sensory stimulation, bioelectronic stimulation or pharmacological perturbation. Our goal here is to change activity in the network either by dampening it or chronically elevating it for a few days. By perturbing the system, we can trigger adaptive homeostatic mechanisms and better understand how the system responds to changes in activity levels.
Jean: Is this dysregulation of activity just a bulk statistics change, or how to you recognize the change or identify what is abnormal?
Sam: We use different approaches to do that. When we overstimulate a neural network, it will cause an acute elevation of cortical activity. Having challenged the system by either removing sensory input or overstimulating it, there is an adaption in the cellular network activity, typically at the population level. Either they'll reduce their activity back to baseline or increase their activity back to baseline depending on the perturbation. So, by challenging the system, we are able to unmask the plasticity process. I like the Ferrari analogy - imagine that you have a Ferrari and you drive it up and down a suburban road. You won't notice its performance capabilities. But if you drive it around a racetrack and really push it right to the limit, then you'll see what it's capable of. That's exactly what we're doing with these neural-circuits, unmasking their hidden processes by pushing the system. If we see dysregulation or failures in homeostatic control then we know the system is unable to adapt appropriately to the perturbation we have given it.
Jean: In your recent nature communications paper, ‘Audio-visual experience strengthens multisensory assemblies in adult mouse visual cortex’, you mention that audio and visual networks are connected by these multi-modal neurons which don’t necessarily network through the thalamus. How complex do you think these sub-networks can be, and do you think it could change our view of the role of executive function?
Sam: So, in the audio-visual plasticity paper, we looked to see if there is evidence for multimodal neurons that respond to both auditory and visual stimuli in primary visual cortex. We found that there were examples of this kind of response and that those neurons showed plasticity when we gave a simple sensory association involving both modalities. The take-home message from that paper is that multi-sensory integration can happen in primary sensory cortices, so at earlier cortical stages than perhaps more traditional models have suggested. Furthermore, we saw strengthening of associations between neurons that was relatively long lasting, implying that networks of neurons can hold a trace of multi-sensory experience.
Jean: Yeah, it's very interesting to have a look at these sub-networks and how they interact and then compare it to this very hierarchical model that people use of the brain, which seems sort of simplistic. Based on all the computational neural network work that's popular right now, how similar do you think the computational neural networks are compared to a biological brain? Do your subnetwork discoveries add computational complexity to our computational neural network models?
Sam: I think that computational networks are incredibly important for neuroscience research. I work very closely with Prof. Claudia Clopath at Imperial, an outstanding computational biologist. Her team are brilliant to work with and working with them has always been incredibly illuminating. Often, discoveries we make in the wet lab can be used to inform computational models, and more often than not, the computational models generated in her lab, have really opened up new testable hypotheses and additional experiments, that we hadn’t previously considered. I think that having this bidirectional relationship with computational biologists and “lab biologists” is really important and helps us to think and crystallize both the experiments that we're doing and the interpretation of the major findings.
Jean: I completely agree. Often, I've seen a separation between the computational neural network models, which have a more simplified neural model, and then the work stemming out of neuroscience, which is finding ever more subtleties in the synaptic processes and the cellular mechanisms. It’s interesting to see a route forward where they join together.
Jean: What do you think are the current challenges in academia? For example, what do you think is slowing down scientific progress?
Sam: That is a really good question. I think that the challenges that we face are due to systems in the different places that we work. So even though some outstanding universities share a very common geographical space, they're likely to have increasingly diverse administrative rules, systems and cultures. That can actually slow down science because it can inhibit the way researchers at these different institutes collaborate with each other. For example, it can be difficult to share resources, reagents or tools. I'm very lucky to be part of the UK DRI, because it is a virtual institute without walls made up of lots of different excellent university organizations throughout the UK. What's fantastic about it is that it attempts to overcome the challenges I’ve laid out and is really built on the core principle of collaboration.
Jean: That's fantastic. It's great that the DRI is doing so much for collaboration across institutions. Coming from a more start up and industry background, I want to ask a commercial translation question about your work. What aspects of your research do you think are most likely to be commercialised?
Sam: Yes, rejuvenation strategies for the aging brain are the most likely thing that will come from our current research. I think now that people live longer, there is an interest in preserving brain health, and not just in the context of neurological disease, which is obviously a massive burden on society. There's more interest in having a healthy, agile mind in later life. I think that we can definitely deliver strategies to achieve that based on some of the work our group and others are doing. One of the things that we know is that as the brain ages, there's a dysregulation of neural activity levels. It creeps up over time. This elevated neural activity impairs neural coding and some cognitive tasks. We know that if you just dampen activity, cognition improves. We know that there's a route to making changes that would improve cognition. A palette of pharmacological interventions perhaps, or maybe even sensory manipulations that you can take to improve your cognitive abilities in later life.
Jean: That would be life changing to so many people if there was a method to improve cognition through sensory manipulations. I was wondering, how has the lockdown affected your approach to research and are there any parts of the life under lockdown that you would consider taking back after we find a way out of the global pandemic?
Sam: As a basic scientist, I think it's probably fair to say that everybody misses the lab. I think I've been very lucky to be incredibly well supported by my team who’ve managed well during lockdown, and managed to be incredibly flexible and wherever possible keep their experiments going. I think that the positives for us are that it gave us a bit of breathing and thinking time to analyse some of the data that we collected in greater detail. We have also had time to reflect and build targeted strategies for the next phase of experiments so that we can really focus our efforts on what the best experiments to deliver the next phase of the experimental program are. It’s obviously been nice to have a bit of time with family at home too, which I think is very, very important for everybody.
Jean: What drives your passion for research?
Sam: I'm not very good when I'm bored, I really enjoy having complex problems to work on. I think understanding complex problems and discovery is a really addictive process. The discovery of something new is a real buzz and is very difficult to describe, even if it's something relatively small. To be discovering how the brain does something is a really privileged position to be in. There's this moment when there's only you and three or so other people who are looking at the data and now know how some element of the brain works. I think it's also very, very addictive because you immediately think of the next question and you immediately want to know how the next thing works. For me, that's what drives the energy in the group. It's like that game, ‘pass-the-parcel’ that you may have played as a kid? When the music stops you get to unwrap a layer of the parcel and then the music starts and you pass it on again, but desperately want to know what's under the next layer. That's what I think science is like. It's always wanting to know what can be discovered under the next layer.
Jean: That's a really nice analogy. I just want to say thank you for taking some time to discuss and answer these questions. Your work seems very interesting. So thanks a lot.
Sam: No problem. It's a real pleasure to talk to you.