 Let's see, tell me when I share my screen. Okay, so I'm going live now and we have started streaming. Okay, so I will start it, yep. Okay, so it's three, so we will start. Welcome everyone to this new seminar of the Sasec Vision Seminar Series. I am Antonio Nojosa, a postdoc in Leone-Lagnado Lab, and I'm working in mouse cortex. And I'm here with Natalie Rochford today and that is gonna tell us about her research on how behavioral context modulates neural activity in the visual cortex. So, sorry. So, first I would like to thank the organizers of the, because this talk is part of the WorldWire Neuro Initiative which was put forward by Tim Bogos and Panos Bocellos. And this is aimed at creating talks that are greener and more accessible to all than the traditional talks. And I would also like to tell you about the structure of the talk. So we will have 45 minutes of the talk and then we will do around 15 minutes of questions. And later on, we will have a more relaxed discussion over Zoom, so please feel free to join if you have any further questions that you would like to discuss with different people about the topics. And also another thing I want to remind is that if you have any questions, please write it in the chat and we will make the questions at the end of the talk. And so, also please join the channel if you don't want to miss any of the talks that we do every week about visual processing. So here, Natalie Rockford is an associate professor at the University of Edinburgh and there she started her group in 2014. So during her career, she started as an undergraduate in where she studied biology and epistemology at the ENS in Paris. And then she obtained a European PhD in neuroscience from the University of Paris 6 and the Ruart University in Wuhan in Germany. She was supervised by Alf Heisel and Chantal Mijeret. She then moved on to do her postdoctoral training at the Technical University in Munich with Arthur Cornet and during her PhD and postdoc, her work had contributed to a new understanding of how neurons acquire their functional properties in the visual cortex. This work also has led to the development of new in vivo two-photon calcium imaging techniques in mouse and her current projects in the lab investigate how behavioral context modulates neural activity in the visual cortex. She's also won various prestigious honors like the Bernard Katz Lecture Award and the Embo Jan Investigator Award. Yes, also she was being awarded the Sir Henry Dale Fellowship and ERC Consolidator Grant more recently. So it's a pleasure for us to have you here, Natalie, and we are looking forward to hearing about your research. Well, thank you very much for this oral introduction. So I will share my screen here, right? Can you see it? Yeah. Yes, I can see. Okay, so today I will present a recent study from the lab investigating how energy availability impacts information coding in neocortex. And before diving into this topic, I would like to give a bit of context explaining how we started investigating such question. So my research group is using the primary visual cortex I need to, yeah, as a model system, as a model of cortical circuit, integrating visual inputs, sensory inputs with contextual signals, such as the motor activity and as well as the past experience, meaning when a stimulus has been associated with the reward, how this association changes the activity in neuronal populations in the primary visual cortex. And we have been studying that at different level, the level of neuronal populations using individual to photon calcium imaging, as well as at the level of single neurons using 3D imaging. And when studying how experience, how learning influence the activity of cortical circuits, we as most of our labs interested in this question, we have used food or water restricted animals that would be motivated. So we are, as a motivation, we are using these food or water restriction and these mice learn rewarded behavioral tasks. And then we are interested in following the changes in cortical activity during the learning of this rewarded task. And while we were doing such experiments, we actually wondered whether these changes in metabolic state associated with this food or water restriction was actually by itself changing the activity of this cortical circuit. So this was a starting point to investigate how a contextual signal linked to a metabolic change would affect the activity of cortical circuits. So starting this study, we know that information processing in the brain is metabolically expensive, despite comprising less than 2% of our body mass. The brain is, the human brain is actually consuming almost 20% of our total calorie intake. And this consumption is largely due to the use of, to the energy spent by neurons to reverse the iron fluxes associated with electrical signaling via the sodium potassium ATPase. So we have a very costly neuronal function and the main usage of this energy is as you can see in these pie charts for synaptic transmission. And despite the fact that the brain is metabolically expensive, the actual resources, the food availability, the intake is highly variable. It was shown, for example, I gave here two examples in the orangutans in South America and in the strip mouse in South Africa, just examples to show how the range of which the food availability and intake can vary across individual seasons and sex here for the orangutans between the males and the females between January and May and here during the dry and the moist season for the stripped mouse. So our question was, how was the mammalian brain adapted to variations in food availability? And given this energetic cost of brain functions and the scarcity of resources, the brain is forced to have evolved an energy efficient coding strategy that would maximize the information transmission per unit energy. And when we speak about energy, it's women ATP. So that would maximize information transmission per ATP usage, unit of ATP usage. Given that this brain is energy limited, one hypothesis is thus that in times of food scarcity, neuronal networks should save energy by reducing this information processing. And there has been previous studies done in invertebrates indicating that that might be the case. So when Clu was coming from a study in a fruit fly where it was shown that dopaminergic neurons that are involved in long-term memory formation would were found to be strongly reduced in their activity in starved flies. So during food restriction, this would strongly reduce the activity of this dopaminergic neurons and this would have a negative impact on this long-term memory score. When the authors exogenously activated these neurons, there was a recovery of the memory score but this was at the expense of survival. So basically the food flies died more. So this was an indication that memory formation is metabolically costly and that these food restrictions would impair this costly long-term memory in flight to promote survival. Another indication is coming from this study from long-term et al showing that food restriction would reduce again in the fly, the gain of the visually evoked neural activity during walking. So here the activity during walking in ground, in stationary, in black and in food deprived animals is gain increase during locomotion would disappear. So again, as an indication that the nutritional state would modulate the neural processing of in that case visual information. But the question of how these changes occurs or which changes occur in the mammalian brain remains largely unexplored. And that's what we wanted to study, how the food availability would impact the energy usage and the information processing in the mammalian cortex. And in other words, we use the model of food restriction of calorie deficit and we were wondering how it would impact the energy usage, would it decrease the energy usage to save energy and this process of saving energy would it be at the cost of information occurring precision in the brain. For this we used as a model system, the primary visual cortex in food restricted mice. Here you see the weight of our control mice in black and our food restricted mice in red. These food restricted mice would lose 15% of their body weight over the course of two to three weeks. And after these two to three weeks of food restriction we would study their neural activity in the primary visual cortex, layer two free using either wall cell recordings or two photon imaging. And very importantly we did that in sated animals because we were not interested in hunger per se or in short-term changes in metabolism but we were interested in the effect of a long-term food restriction over the course of two to three weeks associated with a body weight loss. So the impact of this long-term restriction on the activity of cortical neurons. So in these sated animals we performed first measures to study how food restriction was associated with a potential decrease in energy usage. Again with the hypothesis that in response to this long-term deficit in food intake cortical networks would save energy by reducing the ATP usage. So in order to measure the ATP usage we first used an imaging strategy by using transgenic mice that were provided by our collaborator Johannes Erlinger in Leipzig, Germany. And by using these mice we could use the FRET signal by first using ATP synthesis inhibitors and then monitoring the decay of this FRET signal both in control and food restricted mice. And what we found is that this decay rate was reduced in food restricted mice indicating a reduced ATP usage. These measures were then confirmed by estimating ATP usage from electrophysiological recordings. And for this this estimation is based on the fact that the energy expenditure for neuronal signaling is principally associated with the reversal of the sodium influx via the sodium potassium ATPase for free sodium ion 1 ATP such that if you estimate the sodium influx you can estimate the ATP usage associated to the reversal of the sodium influx. Okay, so this was performed in awake mice and we first saw again this pie chart showing the energy usage and that most of it is done for synaptic transmission and the part for the action potential is associated with action potential. So we first checked in vivo whether the spike rate was affected as well and the ATP usage associated with this spike rate and then we were interested in the synaptic transmission. So first the spike rate by using Wasell recordings in awake mice visually stimulated with natural images we could record the activity of the neurons in layer 2-3 and we found that actually the spike rate was maintained in the control group and in the food restricted group there was a significant difference. As a consequence the ATP usage associated with this spike rate was not different across groups. So we found that under the food restriction there is less energy but despite this reduced energy available the cortex managed to maintain the spike rate. However, when then we were interested in measuring the excitatory synaptic currents using voltage clamp again in the same region in these awake mice we found a decrease in this mean excitatory current corresponding to about 30% of reduction in the ATP usage associated with this excitatory currents and this was showing that one consequence of these food restrictions was ATP savings linked to this reduced excitatory current. We checked whether the inhibitory currents were affected as well and we found that this was the case such that the excitation inhibition ratio was actually maintained between both groups. So in these food restricted mice we have a decrease in the mean excitatory current a decrease in the mean inhibitory current such that excitation inhibition ratio is maintained. Okay, so knowing that we found this decrease in vivo we wanted to go a bit more into the mechanisms and went to in vitro recordings here a slice of visual cortex, mouse visual cortex and we found in this in vitro preparation again that there was indeed a decrease in the EPSC amplitude a decrease in the ampereceptor currents by about 34%. Here in the amplitude, this decreased amplitude again food restrict animals in red and controls in black. This decrease in ampereceptor currents was associated with the decrease in the ampereceptor conductance these recordings were performed with TTX and we could confirm that these changes were due to post-synaptic changes since these ampereceptor mediated miniature synaptic currents were found to have a decreased amplitude but a stable frequency. So there was a decrease amplitude between the control and the food restricted group but the frequency was maintained. There was also no evidence of presynaptic changes in transmitter release based on measures of per pulse ratio and the coefficient of variation of the synaptic responses. So we just found in this first part of the talk associated with the energy usage that the food restriction, this calorie deficit was associated with indeed the reduction in energy usage. In vivo we found this about 30% reduction in ATP usage linked to the decrease in mean excitatory currents and in vitro we found that this was mainly due to this post-synaptic effect. So despite this decrease in the mean excitatory currents in this ampereceptor currents we found surprisingly that the spike rate was maintained as I showed you in the first slide. And so we were wondering how can it be that the spike rate is maintained despite this decrease in excitatory synaptic currents. And I will show you first the mechanism and then show you how we confirm that in vivo recordings. So this preservation of spike rate despite the reduction in ampereceptor currents here you have a model of a neuron here the ampereceptor current, the membrane potential the spike threshold in controls and here in the food restricted animals as I told you there is this decrease in ampereceptor current. What we found and I will show you is that compensatory mechanisms are in place in this food restricted mice such that we found an increase in the membrane potential as well as an increase in the input resistance leading to a maintenance of the spike rate. So by this decrease in ampereceptor current was compensated by the increase in the membrane potential and the increase in the input resistance such that the distance to the spike threshold was compensated and we could have a maintained spike rate. So we confront these mechanisms with the current clamping vivo again in awake mice watching a screen with a natural image is our example of recordings. Here the input resistance that was found to be increased the membrane potential, resting membrane potential that was found to be again increased in the food restricted mice such that the distance to the spike threshold was decreased in the food restricted mice and the spike rate maintained. So this increase in the membrane potential and this increase in the input resistance normalized the spike output when the ampereceptor conductance was reduced and we used a simple model integrate and fire model neuron to show that indeed when we played with this model by decreasing the excitatory synaptic conductance and increasing then the membrane potential and the input resistance we could compensate we could normalize the sub threshold depolarization to the spike threshold such that when you increase these two compensatory variables you compensate the difference between both groups and this was indeed what we found in our in vivo measures that this normalized sub threshold depolarization was indeed normalized across both groups leading to the maintained spike rate. So what I have shown you so far is that during food restriction there is a save of energy due to this decrease in excitatory synaptic current leading to a saving in ATP usage by about 40% and maintained spike rate. In other words it means that these neurons in these situations spike at similar rates as controls but they spend less ATP on underlying excitatory currents. So the question is is there a cost to this strategy because if neurons could spike at the same similar rate as controls but spend less ATP for doing that you could think that this could be a default strategy why isn't it the usual way of functioning for these neurons, right? Because indeed there is a cost and this is a spoiler the cost is the loss of information in the coding precision and that's what I will show you in this second part of the talk. So we found that indeed these compensatory mechanisms ensure that the spike rate was maintained but maintained spike rate doesn't mean that it's the same that there is no that the information is the same. Indeed if we consider that we would with the same number of spikes this is a schematic showing that you could have the same number of spikes but from trial to trial very highly reliable spiking with the spike always at the same time for example for this even grating which would contain a high information compared to a situation where you would have a high trial to try variability with actually the same number of spikes but spiking with much more variability and which would decrease the content of information associated in that case with this visual stimulus and so we went on to test this hypothesis that okay maybe the spike rate was maintained but there was an increase in the noise or in the try to trial variability and we went on to study what would be the consequences of this increase in the try to try variability so we first started by using again a model neuron this Hodgkin and Huxley using one variable input source and testing with this model the advantage of the model is that we could play with the different parameters and compare the impact of increasing the input resistance, increasing the membrane potential and both and what we found again so comparing the control situation with the food restriction food restriction, decrease of unperceptible conductance increase in the membrane potential increase in the input resistance and what we found in this model when we increased these parameters was an increase in the trial to trial variability and when we put both the increase in the input resistance and the membrane and resting the potential we saw a strong increase in this food restriction case compared to control we wanted to know if this was the case in vivo for this we used current clamp in the wake mice watching gratings it is known that in the primary visual cortex of mice you have orientation selective neurons so this is an example of a response of a neuron responding more strongly for these gratings compared to others and we first checked what was indeed the trial to trial reliability of these responses to different oriented gratings what we found so here are some examples of recordings in control in food restricted animals for the preferred angle for the non-preferred angle and what we found is indeed a significant increase in the trial to trial variability in food restricted animals this increase in trial to trial variability was associated with a broadening of the tuning curves so when we plotted a tuning curve so plotting the spike rate in response to the different oriented gratings then we normalize to the preferred response to the preferred orientation plotting the tuning curves and averaging across the control mice and the food restricted mice and you see here a broadening of the orientation tuning curve for the food restricted mice here quantified by the tuning width of these tuning curves that is increased in the food restricted mice with about 30% of broadening in food restriction compared to controls we checked whoever this increase in tuning width so this broadening of the tuning curve was associated with a broadening of the sub threshold depolarization because this could have been one source we said the trial to trial variability could have been one source explaining the broadening of the curve but in principle broadening of the sub threshold activity could also be a source this was not the case you see here the recordings showing that for both control and food restricted animals the normalized sub threshold depolarization was not different from the sub threshold depolarization so this broader orientation tuning in spike output was not directly inherited from the broader sub threshold depolarization but indeed we went further to test where it was associated due to the increased sub threshold variability so again with this idea that we would have similar sub threshold depolarization in both cases but with an increase in the sub threshold variability you would have a broadening on the spiking output just as a schematic representation so in order to test that we went back to our model the Gaussian noise model in which we fed into the model this equivalent sub threshold depolarization but the increase tried to try a variability and what we found is that indeed this so similar sub threshold depolarization increased try to try a variability leading to a broadening of the orientation tuning curve here and you can see here the model and here the in vivo recordings that actually led to very similar results showing this clear broadening of the curve associated with this increased try to try a variability finally we went back to our Hodgkin and Huxley model showing that again when we increased the input resistance increased the resting membrane potential we increased the try to try a variability and leading to a broadening of the orientation tuning curve notably when we removed this variable input source we could of course remove the try to try a variability so there was no more variable input source and this actually normalized the tuning curve such that there was no difference anymore between control and for the restricted noise all together I hope I have shown you that during food restriction less calorie intake we have decrease of energy usage saving energy through for reduction of amperiseptic current amperiseptic current the spike rate is maintained thanks to an increase in the input resistance an increase in the resting membrane potential the consequence meaning an increase in the subfacial variability leading to a broadening of the orientation tuning curves and the final question was whether this broadening of the orientation tuning curves had consequences in the coding precision and impaired perception and to test this we went back to two photon calcium imaging so back to the usual method which can be used in the lab and for this we image neurons labeled with GCAM 6S again awake mice in front of oriented gratings here you have an example neuron responding to these gratings with a preferred orientation this 120 degree we confirmed that the orientation tuning curve was broader for food restricted mice compared to control the tuning width was indeed increased and we tested whether we would also find decreased in coding accuracy when we were showing natural images so we compare the situation where we showed natural images from very different environments so this would be a very easy discrimination compared to since from the same environment which would be a very fine discrimination it's hard to discriminate between these images and when using a maximum likelihood decoder to we when using the activity of the layer 2 free neurons image with the two photon calcium imaging and from this activity we used the decoder to decode which stimulus had been presented to the mice and to infer the discriminability in this course discrimination case and here we find discrimination case and what we found is that for the course discrimination easy discrimination there was no difference between both groups while for the fine discrimination when it was hard to discriminate then we saw a clear deficit in food restricted mice such that the decoding accuracy was significantly lower in the food restricted group indicating again a decrease in this cortical coding precision and we went on to know whether this decrease in coding precision so shown here in the orientation tuning through this broadening of the orientation orientation tuning curve or here with the lower decoding accuracy for fine discrimination whether it was it had some implications for behavior for visual discrimination and for this we used water mace tasks or at least discrimination tasks visual discrimination tasks but done in a big water mace that I will show you a movie yeah so that was adapted with this shape divided here two screens and here you have a platform in front of the target screen with these vertical stripes and the mice are trained to find this platform in front of this target stimulus and once the both groups have been trained and reach a performance of higher than 80% we then tested these mice by decreasing the difference between the target and the non target so at the beginning with a 30 degree difference so it was not too difficult but then by decreasing the difference between these two angles to 2010 5 which made it very difficult to discriminate and then the control with no difference and what we found when we tested the performance of the mice so the percentage of the time in which the hidden platform what we found is a decrease in the performance in these food restricted mice when the angle difference was below 10 degrees so below between 10 and 5 so around 70.5 degrees again indicating that these food restriction results in an impaired fine visual discrimination so what I have shown you so far is that during food restriction we found a decrease in energy usage spike rate was maintained due to an increase in input resistance increase of resting on potential which led to an increase of supression variability a broader tuning curves which decreased the coding precision in response to natural images and had a consequence in as a decrease in the fine visual discrimination as shown in this behavioral test so to now that we I have shown you all these cortical changes associated with food restriction what big question is what's the link between the food restriction or food intake and these cortical changes in other words what's the metabolic signal linking these food restriction to these cortical changes and as a first hint to study that we first checked whether these changes were actually reversible so we had this food restriction this decrease in weight and we wanted to know whether these changes would be reversible when the mice would recover their weight and this is the case so you see here the same plot I had shown you at the beginning of the talk with the weights the normalized weight of the mice the control mice in black the food restricted mice in red here the two to three weeks food restriction period and here after the end of this period the mice had a little bit of access to food and you can see here the recovery of the weight we use two photon calcium imaging again to image their two free neurons before and after the recovery so session one food restricted mice with 85% of their body weight session two recovery when the mice recovered their body weight and you see here the orientation tuning broader for food restricted mice and no difference after recovery and here the quantification in controls no difference before and after and in food restricted mice before and after recovery of their body weight so this was an indication that recovering the body weight recovering the metabolic state would reverse the cortical changes but we wanted to know which metabolic signals were underlying these changes and for this we went to do a full examination of serum levels of blood levels of different markers metabolic markers in order to know which ones would be affected in our protocol of long-term food restriction and here you have the quantification for metabolic markers modulated by short-term society such as the glucose the ketone bodies the corticosterone and the adrenaline and you can see that for these full markers of short-term society the serum levels in food restricted animals used for our recordings were actually very similar to the control levels here again this was done in sated animals after two three weeks of restriction however when we changed we checked metabolic signal that was known to be associated with long-term changes in metabolic states which is the leptin here which is a hormone secreted by adipose tissue and was shown to be involved in long-term society and energy balance we found that actually this level of leptin was strongly decreased in food restricted animals similarly to the decrease in cortical coding precision so knowing that in our food restricted mice again that had lost weight lost weight less adipose tissue less leptin and less coding precision in the cortex and we wanted to know whether if we would restore the level of leptin to control levels we could restore the cortical changes this loss in visual coding precision and the answer is yes when we supplied leptin we used this exogenous leptin supplementation to restore the leptin levels to control levels so here again our control mice this is the weight this is the food restricted animals in red and some food restricted animals that were treated with leptin for one week and here the leptin levels in controls in food restricted mice and in food restricted mice with exogenous leptin supplementation and we then checked before and after leptin or saline administration whether the orientation tuning was changed and what we found is that indeed this exogenous leptin supplementation was restoring the visual coding precision here the tuning with with this oriented gratings no change in controls no change between the food restricted and the food restricted that were given saline solution but here the food restricted animals before and after leptin supplementation with this level recovering the control levels here another measure of this coding precision with the natural images checking the decoding accuracy again in controls food restricted and food restricted plus leptin after the leptin supplementation again showing this recovery of coding precision with the leptin supplementation so altogether I think I have shown you that food restriction in our model of food restriction associated with 15% of body weight loss we found a decrease in serum leptin levels this was associated at the level of cortex with a decrease in energy usage due to a decrease in the ampereceptor currents the concurrent increase of input resistance and increase in resting membrane potential maintain the spike rate but also increase the suppressor variability leading to broader orientation tuning that was associated with a decrease in coding precision and a decrease in the fine visual discrimination and this is the summary abstract showing our general conclusion that I hope or the results I have shown you indicate that the neocortex saves energy by reducing the coding precision during food scarcity we had our food restricted mice the body weight loss associated with this decrease in serum leptin we found a decrease in ATP usage and a decrease in the coding precision that was rescued by the leptin supplementation and the in vivo and in vitro cell recordings could show us what were the mechanisms underlying both this decrease in ATP usage through this decrease in ampereceptor conductors associated with a decrease of sodium influx and a decrease in ATP usage used for this sodium extrusion compensatory mechanisms with the increase in resting membrane potential the increase in input resistance that led to maintenance of the spike rate but also to an increase of suppressor variability leading to this decrease of coding precision so all together our results show that the brain the cortex the neocortex dynamically adjust its energy usage and coding precision depending on the metabolic state so on that I would like to thanks the people who did the work so really the driving force from the beginning to the end was Zaid Badamsi who is a postdoc in the lab and really did all the cell recordings all the in vivo imaging except the ATP one and yes again was really the star of the story the ATP measurements and the behavioral test was performed by Danai Katsanevaki and the model the computational expert was Natalie Dupuy here and I would like also to thanks the funding bodies the Wellcome Trust the EBSRC the Royal Commission and the ERC thank you very much okay thank you so much Natalie it was a great talk really clear and it's a lot of work it's amazing what you've achieved so if you if you like we might go with the questions yes sure so there are many in the in the chat I'll try to go from the first one from Leon Lagnado is the dominant energetic cost of synaptic transmission presynaptic or post-synaptic so we from so we have the evidence from in vitro recordings that it's post-synaptic so I will go back to the side so this was checked again in vitro here yeah so three lines of evidence actually this ampariscepter mediated miniature synaptic currents in TTX so that the amplitude was changed but not the frequency and also that we found that there was no evidence of presynaptic changes in transmitter release based on per pulse ratio and the coefficient of variation of synaptic responses so mainly post-synaptic mainly post-synaptic okay so then we have Henrique Bonn-Helsdorf that is asking if your EPSC's amplitude was reduced so it seems sorry sorry it's not a question it was just a comment in order to Leon sorry but then we have Keith Longden that is saying lovely work did you test temporal properties presumably responses to rapid stimulus changes were particularly affected sorry I didn't understand so if you test temporal properties oh I see so whether there was a delay in the response or did be that she had that comment saying presumably responses to rapid stimulus changes were particularly affected that's a very good point we haven't tested that but that's indeed a very good point we don't know yet so then we have Luisa Ramirez asking how is the reduction percentage of ATP usage related to the calorie deficit amount how how is the reduction percentage of ATP usage related to the calorie deficit amount so our model we use the model of food restriction so we have a food restriction food restricted mice so they have less access to food right they are food restricted they have lost 15% of their body weight and in this model we are taking these mice food restricted mice and we measure the ATP usage with two ways ATP imaging using these transgenic mice that have been used in the lab of professor Healinger and we also estimated the ATP usage by using atrophysiological recordings by estimating the sodium influx so we had these two lines of evidence showing that there was this decrease in ATP usage so food restriction was associated with a decreased ATP usage associated with excitatory current the decrease of excitatory current does it answer the question and we found about a 30% reduction in excitatory current and associated ATP usage yeah I think it does the same person Luisa Ramirez asked if are there effects of surf in different color indefinite amount which is something I was wondering as well yeah so indeed this study we took a bit the extreme case so it's really the model of a long term restriction we haven't studied the effect of short term restriction and we also haven't monitored the time course of these changes so when these changes appear during 2-3 weeks of restriction so we do expect that indeed a short term restriction or just a mild restriction with less body weight loss would lead to different changes or no changes or even beneficial changes so it was shown that mild food restriction that was not associated with body weight loss could actually be beneficial for brain functions okay so we have a follow up from Leon there is a small delay in the chat so it's okay I don't see the chat actually but yeah yeah it's because it's in YouTube so Leon says that the change is post synaptic but he was wondering what about the energetic cost per basically in either condition okay yeah so if there would be a change for the physical release basically I guess so we haven't checked that would be the answer I would think it's unlikely from the results with the minis but we haven't checked okay then we move on and also just to say that we do find well it's not directly but we do find similar result in vitro and in vivo so this decrease in the excitatory synaptic currents is something both both in vivo and in vitro okay okay well maybe we can discuss so then we have Enrique von Gelsdorf this time asking that is there a survival benefit to less code imprecision in food deprived animals during evolution yeah that's a very good point so with the fly study I was showing probably because it's the flies the researchers could push the system to artificially increase the activity of neurons in food restricted animals right to push the system and see the impact on survival and actually the flies died this is challenging in mice both ethically and also experimentally right because for the equivalent we would have to identify which neurons are causing this decrease in coding precision let's say precisely then activate these neurons specifically in the range that would restore the coding precision so that's again a challenge by itself and if when we achieve that check whether this has an impact in the survival of the mice which ethically is not possible knowing that we cannot actually go lower than 80% of their initial body weight so 20% weight loss that said when we did the experiments of leptin supplementation what we found is that in the group for which we were supplementing the leptin we did see that these animals were continuing to lose weight so we had to stop again at this limit but this would be an indication that if you continue to lose weight when you artificially challenge the system this is likely to have an effect on survival ok let's see then we have Silvia I understand that espontaneous favoring rates stay the same during food restriction but do they change in response to visual stimuli or during curves were normalized? sorry this was not clear this is the spiking rate during visualization that is maintained ok so it's actually the spike rate during visual stimulation both during presentation of great things and during natural stimuli ok yes great so well you cannot see the chat but there are many comments saying that it's a great talk but I will show you later it's nice to hear it's nice to hear I think we don't have time for many questions but maybe for a couple more so then from what good do you expect to see if the mice were overfed instead in the case of obesity? maybe they have a blue visual yes yes yes so we thought the same we were like well then if you increase the adipose tissue and increase the leptin maybe we will have a highly visual increase in coding precision so we haven't tested that's the answer it's unlikely to be the case since actually the leptin is highly regulated and in the model of obesity the leptin levels are actually regulated so you would have a down-regulation of the leptin leptin signaling so no we don't expect but we expect basically a saturation point in the coding precision it's unlikely to be improved but we haven't tested ok we have the same question again then from Silvia do you know whether behavioural modulation is changed during food restriction as you saw for invertebrates? yes that's a very good point so actually because we really wanted to go into the cellular mechanisms all our recordings were done in mice in a little tube that were not running that could not run or work they were like in a tube and ok not they could move a bit but they were strongly habituated to this tube and they were very quiet and stationary so we don't have the data for this with this whole data set that said that's something we are investigating right now with previous data sets from the lab so we would be indeed interested in knowing whether this gain increase during locomotion is modified in food restricted mice it could well be in addition to the broadening of the tuning curve that's ongoing work, it's a good point ok we have just one minute I think we have just one question so we may go with it from Maggie through what mechanism could the change in leptin be influencing the cortical processing yes so that's a good point we investigate many underlying mechanisms but there are still many many questions related to the link between this change in this serum levels the leptin hormone and the cortical changes and for this there are actually many different pathways the leptin can pass the blood brain barrier so it could act directly on neurons in the visual cortex it could also be an indirect effect that the leptin would actually cause some changes in over brain areas for example the hypothalamus in itself would affect the cortical changes finally it could be even less direct or more indirect by affecting the levels of over hormones for example thyroid or hormones that themselves then would affect cortical changes so there are many different paths which these changes in leptin level could affect the cortical changes we see we are investigating some of its aspects okay well great so this is it thank you so much Natalie thank you and thank you to the audience for fantastic questions and for all of us to be on time so now we will go over for further discussion in Zoom actually I have already people joining and that's it thank you so much I leave this the chat open for a while and then close it when people join it's because there is a delay so if I close it now they won't see the Zoom link I couldn't see the YouTube right I couldn't see the chat I know we said don't open it if you open it don't put the sound on because then there is a loop and it's messy hello we can hear you you are muted hi Natalie sorry I singlily I singlily failed to get my question across I am sorry actually I saw Simon briefly appear because you know what my question was about is Simon there? yes I am just trying to start my video there we go I think my question might be to you as much as to Natalie the start of your talk you showed this pie chart the energy budget for energy costs in the brain you made the point that synaptic transmission was two thirds of it but of that two thirds how much is the pre-synaptic processes pumping out calcium recycling vesicles filling them up with transmitter versus the post-synaptic cost if you want to increase the energy efficiency of information transmission per vesicle is the best place to do it pre-synaptically or post-synaptically if that makes sense Simon thought about this a lot in our budget which is now really ancient we just took the numbers of ions and molecules that were transferred during signalling and had to be restored and an ion channel lets through something like 100,000 ions or at least a post-synaptic set of ion channels and that's a lot of ions compared with the number of calcium molecules and the number of transmitter molecules so we felt that our calculation showed that post-synaptic costs dominate but a lot is going to depend on how many post-synaptic receptors you have and there may also be costs in the pre-synaptic terminal that haven't been discovered yet another way to answer this there's no mean we are saying that this decrease of amperisceptor current is the only way to save energy that's the only one we tested so we tested the spike rate and this excitatory amperisceptor current but they are very likely over mechanisms that are affected to start with NMDR receptors but also pre-synaptic potentially pre-synaptic mechanisms and also cellular maintenance many other mechanisms are likely to be affected this was an entry point basically you can't do everything yes but you made the point that the mini-frequency was unchanged but I didn't quite get if the number of released vesicles on average was changing so the mini-frequency is unchanged it evoked frequency not spontaneous yes it evoked by the stimulus okay so that's telling you there's no pre-synaptic modulation in your context that's what I thought because the frequency wasn't changed but the amplitude was decreased that would mean that it's mainly post-synaptic I tell you what prompted this question is that we've been looking at modulation in the retina a different context of modulation circadian modulation in zebrafish and there the main uremodulator is dopamine primarily and there are profound pre-synaptic changes in terms of how efficiently a stimulus evokes vesicle release so of course it's a different context but and I don't know if there's any kind of energetic implications of it I guess that's what I was wondering I think the thing you have to think of Leon is that in retinal synapses although you've got