|Home | About | Journals | Submit | Contact Us | Français|
We studied a network of cortical neurons in culture and developed an innovative optical device to stimulate optogenetically a large neuronal population with both spatial and temporal precision. We first describe how to culture primary neurons expressing channelrhodopsin. We then detail the optogenetic setup based on the workings of a fast Digital Light Processing (DLP) projector. The setup is able to stimulate tens to hundreds neurons with independent trains of light pulses that evoked action potentials with high temporal resolution. During photostimulation, network activity was monitored using patch-clamp recordings of up to 4 neurons. The experiment is ideally suited to study recurrent network dynamics or biological processes such as plasticity or homeostasis in a network of neurons when a sub-population is activated by distinct stimuli whose characteristics (correlation, rate, and, size) were finely controlled.
Optogenetics provide a mean to control neuronal activity with millisecond precision. However, neurons are often activated simultaneously either by flashes of light that activate the whole population synchronously or by a light whose intensity is temporally modulated over the whole field of view (Boyden et al., 2005). Yet, several methods exist to modulate the light spatially and have been used to uncage glutamate (Nawrot et al., 2009) or activate channelrhodopsin (ChR2) expressing neurons (Guo et al., 2009) (for review of available methods to stimulate neurons with both spatial and temporal resolution see Anselmi et al., 2015).
To gain spatial control of the stimulation, a first possibility is to use a laser and move its beam quickly over different positions. For example, uncaging glutamate at different dendritic locations has been achieved by deflecting a laser beam with acousto-optic deflectors (Shoham et al., 2005). This strategy is likely viable only if we modulate the light intensity sufficiently slowly over a limited area. Alternatively, a spatial pattern of light can be achieved using phase or intensity light modulators. Holographic technique based on phase modulation permits to obtain an image in three dimensions with a good spatial precision but patterns can be displayed at a rate of only 100 Hz (Papagiakoumou et al., 2010). If a two dimensional pattern is sufficient, intensity modulation can simply be obtained by placing a projector or an array of LEDs in the conjugated plane of the sample (Farah et al., 2007; Guo et al., 2009). This technique has the advantages of being easy to implement, can target many regions of interest simultaneously and has the fastest temporal resolution.
Here we took advantage of a fast video projector based on the workings of a Digital Micromirror Device (DMD). A LED light source is split by an array of micromirrors that can be controlled with sub millisecond precision in order to display any arbitrary pattern of light (Barral and Reyes, 2016). An image of the projector is focalized to the sample plane via a pair of lenses and the microscope objective. The DMD technology offers an unprecedented temporal precision that enables to display patterns at 1.44 kHz and even faster DMDs are now available. In our settings, the resulting pixel size (2.2 × 1.1 μm) was sufficiently small to stimulate single neurons.
To activate a single neuron, we selected a region of interest of ~30 × 30 μm, centered at the soma of the neuron of interest and sent a 5 msec pulse of light. By designing patterns that are projected onto the sample, we could target independently and simultaneously a large number of neurons (10 to 100 neurons). Stimulated neurons were both excitatory and inhibitory (expression of ChR2 under the Synapsin promoter) and were activated by Poisson spike trains. The rate and correlation of the spike stimuli were controlled by the experimenter (see Barral and Reyes, 2016). By recording from neurons that expressed ChR2, we verified that stimulated neurons responded faithfully to the light pulses. We then recorded concurrently the membrane potentials of up to 4 neurons in cell-attached and in whole-cell configurations to isolate the spiking activity and the postsynaptic inputs, respectively.
Jeremie Barral was supported by a Human Frontier Science Program long-term postdoctoral fellowship (LT000132/2012) and by the Bettencourt Schueller Foundation. Alex Reyes was supported by grants from the National Institutes of Health (DC005787-01A1). This protocol was adapted from procedures published in Barral and Reyes (2016).