In 14 mice, cortical stimulation frequency produced two neuronal responses: 8 Hz synchronized spiking across the network, while 140 Hz briefly shifted firing rates near the electrode
In 14 mice, cortical stimulation frequency produced two neuronal responses: 8 Hz synchronized spiking across the network, while 140 Hz briefly shifted firing rates near the electrode
On August 24, the authors posted a bioRxiv preprint on electrical cortical stimulation in 14 awake mice. Two thin silicon Neuropixels probes recorded spikes from 2 725 individual neurons across 53 brain regions, both near the electrode and in distant regions on the same side of the brain.
The same sinusoidal current produced two responses. Across the recorded network, neurons began firing at a specific phase of the waveform, while their mean firing rate changed very little. Near the electrode, some cells showed a change in firing rate at the onset of stimulation. The current frequency determined the combination of responses observed in the experiment.
In the preprint, the authors examine the source of this variation. They applied a smooth sinusoidal current to the cortex at 8, 28, or 140 Hz, using ten ten-second trials for each condition. Recordings were compared while the mice were resting and running on a wheel. The waveform made it possible to determine the point in its cycle at which each neuron fired.
At 8 Hz, synchronization strength was predicted most accurately by the electric field strength in each region. At 140 Hz, the best predictor was the density of local anatomical connections. A model that already accounted for distance from the electrode and field strength explained an additional 21–25 percentage points of the variation in synchronization strength when data on connections between and within regions were added. The authors attribute this difference to a change in the dominant mechanism. The electric field of the slow waveform acts directly on cells, whereas the high-frequency waveform is more often transmitted through synaptic connections.
Trials at 5 μA produced the second, local response. During the first 300 milliseconds of high-frequency stimulation, firing rates increased more often in cells that the authors classified as putatively inhibitory based on the shape of the recorded signal. At 140 Hz, these cells accounted for up to 16%, while at 8 Hz their proportion remained below 1%. By contrast, nearby cells with spike features associated with excitatory neurons more often reduced their firing rates. The authors interpret this combination as a brief activation of local inhibition.
Running altered distributed synchronization in 77% of the regions examined at 8 Hz, in 30% at 28 Hz, and in less than 10% at 140 Hz. The local firing-rate shift depended on running in 15% of regions at 8 and 28 Hz and in 0% at 140 Hz.
In this mouse model, stimulation frequency determined the mechanism and spatial scale of the response, while the animal's behavioral state had a particularly strong effect on distributed synchronization. The authors propose that this map could support protocols in which frequency is selected according to the desired response and the animal's state is taken into account when synchronization is measured.