Visualizing mammalian brain area interactions by dual-axis two-photon calcium imaging

Abstract

Fluorescence Ca(2+) imaging enables large-scale recordings of neural activity, but collective dynamics across mammalian brain regions are generally inaccessible within single fields of view. Here we introduce a two-photon microscope possessing two articulated arms that can simultaneously image two brain areas (∼0.38 mm(2) each), either nearby or distal, using microendoscopes. Concurrent Ca(2+) imaging of ∼100-300 neurons in primary visual cortex (V1) and lateromedial (LM) visual area in behaving mice revealed that the variability in LM neurons' visual responses was strongly dependent on that in V1, suggesting that fluctuations in sensory responses propagate through extended cortical networks.

Figure 1

A two-photon microscope with microendoscopes and two optical axes for imaging two brain areas in awake behaving mice. (a) Schematic of the optical pathway. The beam from an infrared ultrashort-pulsed Ti:sapphire laser is divided into two beams using polarizing beam splitters (PBSC). Rotatable half-wave (λ/2) plates control the power of each beam independently; a beam block (BB) absorbs the unused power. Two pairs of scanning mirrors independently sweep each beam across the two individual specimen planes. Each arm of the microscope has three remotely controlled motorized stages providing three translational degrees of mechanical freedom. Each arm also has two rotational degrees of freedom that are adjusted manually. To allow the two chosen brain areas under view to be either distal or nearby, two microscope objective lenses focus the two beams into a pair of microendoscopes (0.5 NA), which in turn focus the laser beam onto tissue. Visible fluorescence emissions return through the microendoscopes and objective lenses and reflect off dichroic mirrors. Photomultiplier tubes (PMT) detect the fluorescence signals. Inset, the two microendoscopes focus the two beams onto the chosen brain areas and collect fluorescence signals. One microendoscope has a holder that bears a miniature mirror folding its incoming beam by 90°; this arrangement allows the objectives to be placed close to each other without collision. (b) Schematic of the mechanical design. Inset, magnified view of the mechanisms for delivery of the laser beams to the brain areas. The mouse is roughly to scale. Obj., objective.

Figure 2

The dual-axis microscope permits imaging of either distal or nearby brain regions. (a) Simultaneously acquired images of two distal neocortical areas (M1 and S1; ~3.5 mm apart) in an anesthetized mouse expressing tdTomato in parvalbumin interneurons. Inset, schematic of the relative locations of M1 (pink dot) and S1 (yellow dot) in the right cerebral hemisphere. (b) Simultaneously acquired images of two proximal visual cortical areas (V1 and LM; ~1.2 mm separation). The microscope can image broad fields of view (top, 708 µm across) and can also zoom in on neuronal processes (bottom, 264 µm across). Inset, schematic of the locations of V1 (blue dot) and LM (purple dot). Scale bars, 300 µm (a and b, top) and 100 µm (b, bottom).

Figure 3

Ca²⁺ imaging in awake behaving mice reveals that neurons in visual areas V1 and LM exhibit supralinear inter-area correlations. (a) Simultaneous Ca²⁺ imaging of V1 and LM in head-restrained mice at liberty to walk or run on a trackball. Grayscale images are time averages over 5 min of movie data. (b,c) Color images of neuronal cell bodies that showed Ca²⁺ activity (b) and corresponding time traces of Ca²⁺ dynamics for 15 V1 and 15 LM cells (c). Blue shaded epochs (c) denote periods of mouse locomotion. (d) Histograms of the mean rates of Ca²⁺ transients (cell bodies and processes combined) from four mice (864 V1 neurons and 418 LM neurons in total) during periods of mouse locomotor activity and rest. (e) The number of neurons active in LM within 1-s time windows is a supralinear function of the number of pyramidal cells concurrently active in V1, under identical conditions of visual presentation, for each of the four directions of grating movement. Traces’ colors correspond to those of the arrows, which indicate the direction of grating motion. Black traces show averages over all directions of grating movement for each mouse. Red traces show three-parameter fits to the form f (x) = a + bx^c. (f) The supralinear dependence shown in e is present in all seven mice examined. Each black trace is the average over all grating directions for one mouse. The colored traces show the corresponding parametric fits, with each color denoting an individual mouse. Scale bars, 100 µm (bars apply to a and b).