 On Iím talking a bit about my work that fits in directly with the talk that you just heard, dealing with cell types, but I want to start by talking about behaviour, because in the end thatís what weíd like to get to. So my labís interested in trying to understand synaptic mechanisms of sensory perception, we like to get a causal, and mechanistic understanding of just how very, very simple forms of sensory perception arise, a wrth gwrs, sy'n gweithiofa fewn cysylltu'r chwarae i gyffredinol amgyrchumerau deilio y cefnod yn swyddiadau, mae'n ychynig mewn proses. Rydyn ni ydych gyffredinol. Rydyn ni'n ddweud cysylltu'r chwarae i gyd i cysylltu'r hwn ac mae'r cysylltu'r chwarae i gael cysylltu i gael cysylltu y cefnod yn swyddiadau gwneud. ond bydd o gweithio'r bobl gyntaf, ac rydych yn storio'n eu cyfnod i gyda niwroffa gwasanaeth o gwasanaethserau gwneud a'r wneud yn y ddylch yn gweithio'r cywrs hon. Felly mae'n erioed i chi'n gweithio y gweithiannol o'r niwroffa gwasanaethau gwneud y gweithiannol, ac ymdweud ysgol cysylltu llunio'r prosesrwyr. Rwy'n dechrau i'r cynnwys cystafol rhywbeth, ac rwy'n dechrau i ddim yn rhan o'r ffordd. Yn cael ei gweld y dyfodol, y dyfodol, ydy'r pethau'n gweithio'n gwerthfyniadau'r prosesrwyr, ond rwy'n dechrau i'r prosesrwyr, SFK is convenient in terms of having access to genetic tools, and MICE have one whisker system that it is useful to get quantitative understanding of what is going on. Further processed in the primary somatosensory cortex. In addition to having a feed forward sensory pathway there is active motor components, an area of primal motor cortex, bwysig a bwysig yn ddechrau'r byddol yng ngyfnodau yn ddweud sy'n ddweud y cwysgwr arwain. Fe'n ddweud, mae'n rhaid pethau sydd yn ffordd y cyfnodd yn yr unig a phobl iawn, yn ffordd iawn i ddweud eich cyfnodd â ddweud yn ddweud, a mae'n ddweud o'r ddweud o'r ddweud. Ond, mae'n ddweud i'r ddweud o'r ddweud o'r ddweud o'r ddweud, oherwydd rwy'n iechyd yn rhaid i'r systemol, Felly mae'n golygu o wieg o dechrau fel dechrau wedi'u falch. Mae dyfodod o'r rhwylo yn bwrw, a dyfodod o hyd y trafnod mewn bethau'ro yn bwysig. Felly, gallwch yn wneud byth o'r unrhyw bobl maes o'r bwysig gyda'r ffinafol a'r maesau. Efallai, y dyfod y ddechrau i'r maes, gyda'r maes iTrsyn yn 1986 o ran rhai. Mae'n tyfu i'r maes yn argyflogol yw mynd i gyrwych â cyd-foclwyr yma, a sut nid i bethau ymwneud eich platfform ar gyfer hynny, a ydych chi'n gorfod cymaint ar y blaenau yma? Mae'r cerd-fodol ar y ffordd yn y platfform yma a'r pwgiau cyfwyr yn ymwneud yma dywed y gallu ymddangos ei pynd ikergynno'r platform fel yna'r pwgiau. Ac mae'n ymdyn nhw, o ran yr ysbyt yw, yn y platfform yr endo'r platfform a pwgio'n gwneud i ymdyn nhw'r pwyllteid. ac roedd yn ystod gweithio ac yn trwyddo. Yr unrhyw ddweud yma'n mynd i gweithio ymweld yma yma ymarfer o'r ffilm sydd yn ymlaen. Ymryd o'r gweithio'r cyfrifolau'r cyfrifolau'r cyfrifolau'r cyfrifolau yma sy'n gallu gweithio'r cyfrifolau ar y platfoedd yma i ymlaen i'r tynnu. Ymlaen oedd y cyfrifolau ar y platfoedd yma, ymlaen o'r cyfrifolau ar y platfoedd, yw'r cyfrifolau, yw ddweud yw'r cyfrifolau, ac ydy'r ddweud y cyfnod i gynhalu. I chyddo i'n meddwl bod dwi'n gymryd yn gweithio'r proses. Mae'r cyfnod yn solygu'r gyda'r hollu. Mae'n ynnig i'w blwyddyn yn ymgyrch i'r cyfnod y targfod ac yn gweld o'r cyfrifio'r gwahodd y tîm. O'r gweithiau yn ei gwirio'r celf sydd wedi bod yma bod ymlaen i celfwyrwyr senfritio'r gwnaeth o fyddech chi'w o blighwyr aio wedi'u gwasanaeth bai hwn o'r gyfer dweud o'r swydd ac mae'r ddefnyddio'r cyflogau ac yn bwysig, mae'n rhaid i'r cyflogiau g yr teini newydd yn rhaid o'r cyflogau rhai apwyr. Felly y mwyaf ar hyn o bobl yma yma yma ar ôl pryd ar yr oedd wedi'i gweld yn gweld yn gweithio'r celfwyrwyr a'r gennymau ychydig i ddweud yma'r lefnodau cymrydau a'r gyflawni ar gyfer y Llyfrgellau Gwyrdol, a phos ymgyrchai'r mynd i'ch bod gafodd oherwydd i gyfer eich gwirioneddau gyda'r gyfer gyfoedd iawn. Mae'n gofynu'r trafodaeth yma o'r gofynu'r Llyfrgellau Gwyrdol yn fyrwgellau'r cael eu gwirioneddau. Felly, felly, yma'n gweithio yma yma o'r lefnodau ar gyfer y Llyfrgellau Gwyrdol, yn y cyfnod mwy o'r cortheithau mwy o'r cyfnod, sy'n ddiddordeb llyfrwyr arall, rydych yn ddiddordeb llyfrwyr ar y cyfnod o'r cyfnod o'r cyfnod o'r cyffredinol, y Lair I, y Lair II, y Lair III, y Lair IV, y Lair V, y Lair V, y Lair V, sy'n ddiddordeb llyfrwyr ar y cyfnod. Mae'r wneud i gynnwys ar hyn o'r nôl i'r fawr yn i'r fawr, ac ydych chi'n gawr yn cael ei wneud i'r gawr yn cael eu newon, ac mae'n gwneud yn cael eu gawr yn cael eu 20% yn y ddisgrifennu cymdeithasol. Ac mae'r fawr yn cael eu ddifol o'r newon, yn fawr, yn cael eu fawr yn cael eu newon, fel i'r fawr yn cael eu cyllidau fawr yn cael eu ddifol, a'n iawn i'n amgylchedau ei ddechrau yn cael eu ddifol i ddifol. A coedd y dddw hospital o ddylch yn gwneud llwyddoedd y mys ond rhan o'ch gynedigydd ac mae'r genedigau yn y mawr yn edrych yn bobl canfaenon fod yn dyn nhw'n hals honno flwrs yw'n hefyd, ar gyfle cyffredin gydy'r cyffredin honno, o'ch cyffredin cyffredin cyllid a rechidol yn gwaith, ac yma efallai ar y cyffredin cyffredin cyllid, gyda'r gyfrigurion Lugwyr yn y maes bryd grŵn. Yn ymweld, yn ymweld, mae'n ddau'r cyfrigurion ar y maes bryd. Ond mae'n ddweud yn ymdweud yn y gweithio'r cyfrigurion gwahanol. Mae'n ddweud yma, mae'n ddweud yma, a'n ddweud yn y maes bryd yn y gyfrigurion gwahanol. Mae'n ddweud yn ymdweud, ac mae'n ddweud yn y gweithio'r gyfrigurion gwahanol, mae'n ddweud yn y gaf 67gfmau. A rwy'n edrych yn fawr i gael'r wneud o ddegwyddiadau i gyflwynoedd diolcholion o'r Lluwpxiontau yn adroddodd ar ddegwyddiadau a'r ddegwyddiadau ar gyfer yn yma, yn oedau yn gweithio'r dreuniau sydd yn cymryd i'r dreuniau. A mae'n gyflwynoedd o'r brain yn oeddau yn credu cael eu cyfrannol a rwy'n mynd i gyd-degwyd two much. It moves on this order of tens of�� op, of군s on maybe five microns or so, depending a little bit on how active the mouse is and it turns out that that type of movement artifact doesn't interfere dramatically with the whole России recording technique that I want to show you and that we have within high quality membrane potential recordings from these neurons. Now gave the urgent neurons, our device sort of a i'r sefydliad cyflwyno yng Nghyrgrifol, a rydym yn fawr ddaf yn gallu y Wirewyr Caiwgwch Reoli, te体oliad ddim yn gweld ff virus, i'r grwp c moes i'n sefydliad i'r grwp hwnnw'r Gaba Ergwyr Nghyrgrifol. Yn y gallwch yn gweld Gaba Ergwyr Nghyrgrifol yng Nghyrgrifol, sydd heb gêmwyr piwydig wrth gwneud gael Gheiddiogol i fod yn ei ddweud gyda'r newons gwneud mewn gwirol iawn, i byddwch ar gyfer y bwyd drwsio, ac rwy'n ddweud rwy'n dweud â siaradwyr Cymru yn yr un yn ddeni'r gwirol iawn. Yn rwy'n dod o'r parwapau a'r gwirol iawn yn gyfreilor na anghydd cynyodaeth a'r yn rhyngwly kolygowodol iawn 5-HT 3a, y cwmbrachau'r gwirol iawnr ar gyfer y gwirol iawn, mae'n defnyddio', yn gweiniogol, yn cael bod rhywbeth wedi gwirol iawn. Roedd yna y mae'n meddwl yn seithio heimerciaeth and compare with a total number of Gabaergic neurons, they find that they can more or less account for all the numbers of neurons with very little overlap in this classification scheme. We think that this is one that's interesting and appropriate and also quantitatively it just fits in well with the cell types and cell numbers that we found in primarysem Psych gern予s of their wake behaving mouse. In our recordings that we used the two-photo microscope, We see our green fluorescent cells. We can use different types of genetically engineered mice to differentiate them, and we find the same cell types as Godfisher and Bernado Rudi identified. There are somatosstaten expressing neurons, fast-pigging, labourergic neurons that expressed parvalbumin, non-fast-pigging, labourergic neurons that don't express parvalbumin or somatosstaten. They fall into roughly this subset as a division of the labourergic neurons inside layer 2, 3. a dyna maen nhw'n gwneud i'w rwyntio'r cyffredinol y llunio niogwyddych, a gwneud i'n wneud i'ch gyd yn y ddechrau. Ac mae'n pethau ychydig o'r problemol yma, dwi'n gwaith yma y tuffordi gefnogi, oedd eich cyd-io'r gwneud i'ch nhw'n gweithio'r ddifrwyng. A mae'n gweithio'r gwneud o'r pethau o'r gwneud eich mynedd. Felly, gwneud yn yn ffasbwynt i'r gyfer yn gyfnodig gyda'r gwneurolydau, N�heddech yw'r bydd n ending Diolch�, hwn ydy'r unrhyw gan yau hwnnw yw'r fHS, mae o hyd yn o��� yng Ngheithnogaeth scaf yn gyflow peathent, ond maeinäu pili Instatol Ap but man o野f 5, r phrall g呼b sydd cou nicely, i buddillueu gallu othrygu am y droi initiallyny ar GymrydDP fuel ar Carryabordなんだ ac efo'r Larygiynaeth fwrdd cyhoedliannos gwelwyr. Yn meddwl beth yng Ngheith � posbsawd y nautyaig, áll yng Nwr Newhyn Pwleg Menhaf a Gweithringsenciaeth penderf котораяan, mae gwaith o'r cyffin cynllun yn ymlaenio ychydig o'i niwhapodau sp Hyun, a drwy'r cyffin cyffin yma yn y mharistwyd gyfer gyffin cyffin cyffin cyffin cyffin cyffin ar gyfer gyffin cyffin cyffin cyffin ymlaenio oherwydd ein cynnig yn ymarfer rydych chi'n olygu o'r cyffin cyffin cyffin cyffin cyffin cyffin cyffin cyffin cyffin cyffin cyffin cyffin a llunio'n cyffin cyffin yn ymlaenio. Felly mae hynny'n gael ei hyffroud이 gyda llunion. I think that's sort of obvious. They also have differences in terms of their just standard electrophysiological properties. The input resistance, for example, of the somatostatin cells is much higher than for the other types of neurons. Consequently, they also have a much lower real base so they're very sensitive to injection of small currents and they'll readily fire action potentials under many conditions. So this is sort of just the basic intrinsic properties of these neurons. A nice thing, of course, about recording in vivo is that there are spontaneous membrane potential fluctuations that may be relevant to the animal's behavior. So in the first instance we look and see what happens in the animal during what state that we call quiet wakefulness where the animal is not moving. We do high speed video filming of the animal and in particular we film the position of the whiskers and when the whisker isn't moving then we call that quiet wakefulness. And during quiet wakefulness there are slow membrane potential fluctuations in basically all types of neurons that we recorded from. So excited neurons have large amplitude membrane potential fluctuations. So here's the scale bar, 20 millivolts, one second. So these are slow changes in membrane potential. There are of course also fast changes in membrane potential but there are these slow envelopes that are obvious in excited neurons in non-fast-biking gap, ergic neurons, fast-biking gap, ergic neurons and the somatostatin cells are unusual. They seem to lack these slow oscillations or at least they're much smaller than in any of the other cell types. You can also see this if you look at the spectrograph here in the Fourier transform of the membrane potential. You see a large peak here for the fast bikers, non-fast bikers and excited neurons. They have a lot of power here in the slow frequencies and that's about half in the somatostatin expressing gap, ergic neurons. So the somatostatin neurons seem to have unique membrane potential dynamics that all the other types of neurons have been recorded in the superficial layers of the neocortex. They're also the most depolarised neurons, all the ones we've recorded from, more depolarised than the other gap-ergic neurons that in themselves are much more depolarised than excitatory neurons and both somatostatin and fast-biking gap-ergic neurons and to a small extent the non-fast-biking gap-ergic neurons are firing spontaneously quite high rates of action potentials which is different from the excitatory neurons which are almost silent under these conditions. They have a median spike rate of 0.1 hertz under conditions that are quite weak from this. So that's all very well. We see the membrane potential of individual neurons here. Of course we would also like to relate that then to the neighbours and so Luke Jean-Tier, when he was in my lab, began doing double recordings from nearby neurons again in the awake head restrained mouse and here he's recording from two excitatory neurons and you see again the slow membrane potential fluctuations that are characteristic of quite wakefulness and the grey neuron and the black neuron have almost identical sub-freshhold membrane potential fluctuations but action potential firing turns out to be uncorrelated. So membrane potentials of all neurons let's say as a first order approximation all go together so the whole neocortex under quite wakefulness is swinging up and down by these huge changes in membrane potential so again 20 millivolts, these are massive swings in membrane potential and all the excitatory neurons in the superficial layers are more or less doing the same thing. The same is also true of the GABAergic neurons. You take a fast biker and you compare it to the excitatory neuron recorded nearby and you'll again see very high correlations in membrane potential fluctuations between GABAergic neurons and the glutamatergic pyramidal cells nearby. And you can do the same thing for non-fast biker neurons, you can take fast biker, non-fast biker neurons, you can take any combination you want and we always see high correlations in these slow membrane potentials during quite wakefulness. On the other hand, if you look at the somatostatin neurons you see that things are a little bit different. First of all of course I told you that the actual slow membrane potential fluctuations are already much less in these neurons than in any of the others but if one then looks more carefully and you start doing cross correlations you'll see that actually there's a small weak anti-correlation where the hyperpolarisation in somatostatin cells follows excitation in the other cell types and so what we think is happening is that the GABAergic neurons will depolarise at the same time as the excitatory neurons and probably they'll drive an inhibition of the somatostatin neurons at the time where the other neurons are receiving excitation. In addition, the somatostatin neurons might lack one of the important excitatory inputs at all the other cell types. So all the other cell types, if it's an excitatory to an excitatory connection, excitatory to fast biking, excitatory to non-fast biking, single spike is very nicely translated to an excitatory post-synaptic potential so there's a nice, as a word, a high release probability of these synapses. On the other hand, in vitro work that's been done by a variety of labs like that Sackman's laboratory, Barry Connor's laboratory, Henry Markham's laboratory has shown that the excitatory synapse to the somatostatin neuron or the martinati cell as it's also sometimes called is a facilitating synapse so a single spike may only give a very small EPSP and so it could be that another reason here for, as we were seeing, no correlation is that it also actually lacks the major excitatory input that all the other cells are seeing around it and in one particular example, Luke Shonte was lucky enough to get a connected pair of neurons in awake animals so we think this is the first case of a synaptic connection studied in an awake mouse. So here's our excitatory neuron, here's a somatostatin neuron and when current is injected into the excitatory neuron Luke noticed that the somatostatin neuron would start firing action potentials by itself and you could do this repeatedly and here you see that you can actually drive the somatostatin neurons to quite high firing rates. When you then hyperpolarise the somatostatin cell you have to remember these are very excitable, these neurons so you can easily hyperpolarise them with just small currents so here injects 100 picoam pair, firing is blocked in the somatostatin cell it continues injecting currents into the excitatory neuron firing action potentials and you'll see that these early action potentials these first one, two, three, four action potentials actually don't give rise to an EPSP it's only the fifth spike here that gives rise to this quite large input and presumably as you keep going with these further inputs these are the ones that in the end were driving spikes under the normal condition and so I think it's entirely feasible to think that in many cases one reason that the somatostatin cells are out of phase and out of sync with all the others is that actually they simply lack excitatory input because the firing rates in the excitatory neurons turns out to be very low and the median rates at 0.1 hertz so it's very rare to get a huge spike train like this from an excitatory neuron occurring and so we think that's one reason that there may be this situation that you have an anti-correlation that's present in these somatostatin neurons different from all the other neighbours that are highly correlated in their membrane potential fluctuations so that's a little bit about the spontaneous activity of these neurons during quiet wakefulness but of course we normally think that this is a primary sensory area it should be involved in sensory processing and so we're also interested to see what happens when the whiskers are deflected and in this case what we do as an experimenter we take the whisker and we simply deflect it we move it briefly for one millisecond we give it a little jolt and this then activates the sensory pathway that I discussed and we call this a passive whisker stimulus in the sense that we deliver the stimulus it's not the animal that's decided to get the information we delivered it so at this dotted line here with a small bit of piezo that gives us one millisecond deflection of the whisker and all neurons basically depolarise so an excitatory neuron will depolarise non-fast bikers they depolarise fast bikers they depolarise the fast bikers are the ones that respond best in terms of getting action potential output then come the non-fast bikers and the excitatory neurons are fairly poor in terms of their response probability many of them don't fire spikes a few of them do and that then averages out to giving you a small brief amount of firing once again the somatostatin neurons these potential martinati like neurons are unusual you stimulate and this is the only type of neuron that we found that hyperpolarises in response to sensory stimulation we've never seen this before we've been recording for over a decade in this brain area and we've never found neurons that hyperpolarise in response to sensory stimulation until we came to these somatostatin neurons they also decrease firing rates so they're spontaneously quite active firing around 5 hertz you stimulate and they stop briefly for a brief period of time and they do this robustly so somatostatin neurons are also unique in the sense that they're inhibited by sensory stimuli quite different to all the other neurons that are excited and fire higher action potential firing rates finally I started off by telling you that the main way that the mouse acquires sensory information is when it actively moves its whiskers to touch objects and so the first thing one might want to look at is to see what happens when the animal is moving its whisker and not contacting an object but at least it's an active process so here we do high speed video filming we see what the animal is doing with its whisker here it's sitting still for long periods of time this is this period of quiet wakefulness and then every now and then it'll start moving its whiskers it's interested somehow in its environment there's a major change in brain state and one thing that happens are the somatostatin neurons hyperpolarise, they turn off the animal stops moving its whiskers there's a little bit of jitter around here and it's whisker, there's a little bit of hyperpolarisation here the animals calm again and their action potentials have been fired now we introduce an object these grey bars here so we put an object in the way so now when the animal is moving its whisker it's actually touching something this I guess is the situation but the animal normally gets its major impact we know in excite neurons fastbiking neurons and non fastbikers that active touch as we call it when the whisker contacts an object induces a depolarising sensory response here there's no obvious sign of depolarisation occurring if you look here on sort of the average trace you'll see that in fact there's a small hyperpolarisation here also at least the somatostatin neurons are clearly not getting interested in active touch once again they turn off and so the somatostatin neurons seem to be unique and different from all the other nearby neurons they hyperpolarise and reduce action potential firing whenever anything interesting is going on in this brain area whenever there's a sensory input whenever there's an active bout of whisking or when the animal is actively touching something so these are neurons that seem to turn off when there's something interesting going on in that brain area now somatostatin neurons are interesting in the sense that they have an axonal projection that innovates layer 1 and so whenever these somatostatin neurons are firing action potentials they're releasing GABA up here in layer 1 and so you might imagine there could be interesting processes going on in the layer 1 dendrites when we turn off inhibition if you turn off distal inhibition you'd expect maybe that the distal dendrites might become more active and so we had a look using genetically encoded calcium indicator GCAM3 we express it in excitatory neurons and when the animal is sitting quiet and we look in the layer 1 dendrites here there's very little going on when the animal moves its whiskers we see dendrites but I don't know if you can see the splines but it does light up and it lights up robustly so apparently as these somatostatin neurons turn off as the animal is moving its whiskers we release inhibition and it looks like the distal dendrites have become more active and start maybe firing calcium spikes or NMDA spikes so there are many things that happen in the circuits during whisking and so we wanted to have a more direct check also on what's happening and so we applied optogenetics we express haloradopsin 3.0 we now also switch over to Cree driver mouse line from Taniguchi and when we turn on the yellow light there's a hyperpolarised somatostatin cells and we can see a complete shut off of action central firing so the optogenetics works nicely and the interesting thing of course is to record in the nearby excitatory neurons when you turn off the somatostatin cells the slow oscillations continue to change in brain state but we do see that there's getting these multiple spike bursts that we think are characteristic of distal dendritic input that turns out to be slow and that could be then driving the calcium signals that we saw previously so the functional role we think now turning back to sort of the big picture where the animal is moving around exploring its environment there must be motor signals in motor cortex motor cortex sends profound projections to layer 1 of primary somatosensory cortex and also deeper layers these somatostatin neurons look like they're well placed to control this part of the input they can decide whether this layer 1 input from motor cortex should be playing an interesting role or not when the animal is actively exploring somatostatin cells turn off and they may then allow and gate this motor input into primed somatosensory cortex and that might then be a critical role in terms of establishing sensory motor integration in other sensory cortices it may be that these somatostatin neurons are involved in just regulating in general top-down feedback we think of this M1 signal as a top-down signal and many other brain areas receive top-down input into M1 and so it could be that somatostatin cells are playing a similar role in other brain areas I want to finish here by thanking Luke Zhongte who is a guy who did all the whole cell recordings Eve Cramer did the calcium imaging Hiroki Taniguchi and Josh Wang from Colesping Harbour generated and Jochen Steiger helped us with immunistic chemistry and I'd like to thank you for your attention We can take a few short questions Betty I think you said something and then just kept on going about the spikes and it's clear that the the membrane potentials are very correlated Did I hear you correctly that the spikes are not? Yeah, so that's absolutely true so maybe just to specify quickly excitatory to excitatory neurons so spike rates are low so it's not that we have really great data on this but if we take the spike data that we have there's no correlation it's just a flat line that the spike rates are that's not true for the excitatory versus GABAergic neurons or GABAergic their spike rates are much higher and they more or less follow the envelope of these slow membrane potential fluctuations but for the excitatory neurons the slow membrane potential is not enough to bring it to spike threshold so the excitatory neurons are hyperporized by about 10 millivolts the slow oscillation has about the same amplitude and so they still need another 10 millivolts to hit spike thresholds spike thresholds pretty much the same in all these cells and so we think there's an additional thing that needs to come in to follow the excitatory neurons and it looks like they need an additional big 10 millivolts input to arrive Okay Any further questions? We are very much convinced Thank you very much