iGluSnFR wide-field imaging responses to 2-72 Hz flicker

Human primary visual cortex (V1) responds more strongly, or resonates, when exposed to ~10, ~15-20, ~40-50 Hz rhythmic flickering light. Full-field flicker also evokes perception of hallucinatory geometric patterns, which mathematical models explain as standing-wave formations emerging from periodic forcing at resonant frequencies of the simulated neural network. However, empirical evidence for such flicker-induced standing waves in the visual cortex was missing. We recorded cortical responses to flicker in awake mice using high spatial resolution widefield imaging in combination with high temporal resolution glutamate-sensing fluorescent reporter (iGluSnFR). The temporal frequency tuning curves in the mouse V1 were similar to those observed in humans, showing a banded structure with multiple resonance peaks (8 Hz, 15 Hz, 33 Hz). Spatially, all flicker frequencies evoked responses in V1 corresponding to retinotopic stimulus location, but some evoked additional peaks. These flicker-induced cortical patterns displayed standing wave characteristics and matched linear wave equation solutions in an area restricted to the visual cortex. Taken together, the interaction of periodic traveling waves with cortical area boundaries leads to spatiotemporal activity patterns that may affect perception.

Methods
In vivo glutamate imaging was performed in awake head-fixed animals using a charge-coupled device (CCD) camera. To capture the fast glutamate sensor iGluSnFR kinetics, images of iGluSnFR activity were captured at 150 Hz sampling rate. The glutamate fluorescent indicator was excited using blue LED light (Luxeon K2, 470 nm), which was band-pass filtered using an optical filter (Chroma Technology Corp, 467-499 nm). Light emitted from excited fluorescent indicators was passed through a band-pass optical filter (Chroma, 510 to 550 nm; Semrock, New York, NY) and a macroscope composed of front-to-front optical lenses. The focal length of the lenses was adjusted such that the field of view was 8.6 × 8.6 mm (128 × 128 pixels, with 67 μm per pixel). To minimize the effect of hemodynamic signal originating from large cortical blood vessels, we focused the optical lens at ~1 mm depth.

Steady-state visual evoked potentials. We used a custom-built setup of white light-emitting diodes (LEDs; luminous intensity 6900 mcd, color temperature 9000K, Model C513A-WSN, Cree Inc.) to deliver visual stimulation (the setup previously used in Gulbinaite et al. 2019 NeuroImage). The LED array was placed in a cardboard box, with a circular aperture subtending 35° of visual angle covered by a sheet of tracing paper (6 cm away from LED array). The visual stimulation box was placed on the left side of the animal, 8 cm from the left eye, at a 50° angle to the animal’s body axis. Flicker was implemented as a sine-wave modulation of the power supply to the LEDs (luminance changes from 0 to 215 cd/m2). On each trial, the animals were exposed to 10 s of flicker followed by 10 s inter-trial interval of complete darkness (Figure 1B). Mice were exposed to flicker frequencies that ranged from 2 to 72 Hz (logarithmically spaced 30 different frequencies). For all frequencies, the stimulation started and ended with maximal luminance (215 cd/m2, π/2 phase of the sine-wave cycle). Each imaging session lasted up to 30 minutes, during which each flicker frequency was presented twice in pseudo-random order. Each animal underwent 5 experimental sessions (i.e. 5 recordings on separate days). Thus in total, each flicker frequency was presented 10 times. 

Visual evoked-potentials. In addition to sine-wave stimulation, we recorded responses to single light pulses (20 ms) using 1500 cd/m2 luminance stimuli to characterize spatiotemporal dynamics in response to single stimuli and on/off responses.