Integrated Calendar

Understanding the neural basis of behavior requires studying the activity of neural networks. Within a neural network, single neurons can have different firing properties, different neural codes and different synaptic counterparts. Therefore, it will be useful to readout from the brain and control it at a single-cell resolution. However, until recently, single cell readout and control in the brain were not feasible. The first scientific problem we addressed, is this regard, was the low spatial resolution of light based neural activation. Opsins are genetically encoded light switches for neurons that cause neural firing, or inhibition, when illuminated (and are therefore called “opto-genetic” molecules). However, optogenetic experiments are biased by ‘crosstalk’: the accidental stimulation of dozens of cells other than the cell of interest during neuron photostimulation. This is caused by expression of optogenetic molecules through the entirety of the cells, from the round cell body (“soma”) to the elongated neural processes. Our solution was molecular-focusing: by limiting the powerful opsin CoChR to the cell body of the neuron, we discovered that we could excite the cell body of interest alone. This molecule, termed “somatic-CoChR” was stimulated with state of the art holographic stimulation to enable millisecond temporal control which can emulate actual brain activity. Thus, we achieved for the first time single cell optogenteic stimulation at sub millisecond temporal precision. A second challenge was imaging the activity of multiple cells at a single cell resolution. The most popular neural activity indicator is the genetically encoded calcium sensor GCaMP, due to its optical brightness and high sensitivity. However, the fluorescent signal originating from a cell body is contaminated with multiple other fluorescent signals that originate from neurites of neighboring cells. This leads to a variety of artifacts including non-physiological correlation between cells and an impaired ability to distinguish between signals coming from different cells. To solve this, we made a cell body-targeted GCaMP. We screened over 30 different targeting motifs for somatic localization of GCaMP, and termed the best one, in terms of somatic localization, “SomaGCaMP”. This molecule was tested in live mice and zebrafish and can report the activity of thousands of neurons at a single cell resolution. A third challenge was voltage imaging in the brain, since genetically encoded indicators still suffered from either low sensitivity, or from low brightness. To record voltage, we used nitrogen vacancy nanodiamonds, known to be both very bright and sensitive to electric fields. Our aim was to bring the nanodiamonds to the membrane so the large electric field created by the action potential could impinge upon them and change their fluorescence. By making the nanodiamonds hydrophobic through surface chemistry modification, and inserting them into micelles, we labeled neural membranes with monodisperse diamonds for hours. We are now in the process of assessing the sensitivity of the nanodiamonds to the membrane voltage.
Altogether, thinking backwards from fundamental limitations in neuroscience is instrumental in deriving strategies to fix these limitations and study the brain. In the future, we will use similar approaches to study and heal brain disease, at single-cell and subcellular resolutions.

Behavioral states, such as arousal and attention are defined by a set of psychological and physiological variables. They have profound effects on sensation, perception, learning, and cognition. In the brain, global states are characterized by distinct cortical and hippocampal EEG patterns. These changes that are clearly observed in the local field potential (LFP) are also evident in network and cellular dynamics. At the population level, the more active states are manifested as asynchronous neuronal firing between neighboring cells. At the cellular level, the membrane potential during active states is characterized by a continuous depolarized state, high synaptic ac! tivity, reduced variance and reduced membrane potential correlations between cells. In recent years it has been demonstrated in rodents that pupil size is a robust indicator of a range of neural activity from neuromodulator release to cortical neuronal membrane potential.
There has been some debate in the field regarding to what extent the effects of locomotion on cortical dynamics are due to arousal and what can be attributed to locomotion. Furthermore, in some studies cortical dynamics were evaluated while the animals transitioned spontaneously between states and in others, states of arousal were externally induced. Additionally, different effects have been reported in the auditory and visual cortex. Therefore, we wanted to more finely differentiate between different states and evaluate the effect of state on the somatosensory cortex.
To accomplish this we conducted intracellular recordings in the barrel cortex as well as extracellular LFP recordings in Layer IV of the barrel cortex in awake head fixed mice. We monitored pupil size as an indicator for state of arousal as well as tracking locomotion.
We found that there is a significant correlation between membrane potential of cells in barrel cortex and pupil size. Neurons were significantly more depolarized as the animal was in a greater state of arousal. This change was not affected by the mode of inducement of arousal, be it a spontaneous transition into a state of arousal or one externally induced. However, the effect was abolished by the occurrence of locomotion.
We also found that responses to sensory stimuli are increased during a state of arousal but not in a state of hyper-arousal. Inducing the state externally minimized this effect and if the animal is locomoting then the increase in sensory responses is abolished.
We further found that when the animal is in a greater state of arousal there is less synchronization as indicated by the decrease in correlation between membrane potential and LFP. Even more startlingly, we found that the polarity of the cross-correlation was reversed during hyperarousal. This would strongly suggest a reorganization of the laminar network across different states.