Different layers in mammalian cortex have specific projections and circuit dynamics^1-5. In particular, infragranular layers is sub-divided into L5a, L5b, L6a and L6b, which have different connections. For instance, L5b and L6b have corticothalamic^3,4 and horizontal^3,5 connections, respectively. It can be tacitly assumed that these layers have distinctive circuit dynamics^1-2. However, how stimulation frequencies at different infragranular layers contribute to widespread and large-scale cortical and subcortical activities remains poorly understood. Furthermore, majority of previous studies utilized electrical stimulation which excites non-specific neuronal populations. In light of these issues, synergistic use of optogenetic and fMRI enables mapping of such large-scale neuronal activity with a cell-type specific, millisecond resolution and reversible neuromodulation technique⁶. Hence, our optogenetic fMRI (ofMRI) study aims to investigate layer and frequency dependent activities by driving excitatory neurons in different infragranular layers of visual cortex.

Animal Preparation and Optogenetic Stimulation: AAV5-CaMKIIa::ChR2(H134R)-mCherry was injected into infragranular layers of primary visual cortex of adult male rats (n=10) (Fig. 1a). After 4 weeks, an optical fiber was implanted to the superficial infragranular layers (L5b, n=5) or the deep infragranular layers (L6b, n=5) (Figs. 1b-1c). Blue light (50mW/mm2) at 1Hz (100ms pulse) and 10Hz (30ms pulse) was delivered for stimulation in a block design. Opaque tape was used to ensure no light leakage causing visual stimulus. Naïve animals (n=2) were used as controls.

fMRI protocol and Data Analysis: fMRI data was acquired at 7T using GE-EPI (FOV=32×32mm2, matrix=64×64, α=56°, TE/TR=20/1000ms, 12 1mm-slices). Standard preprocessing followed by GLM analysis was applied to identify and extract significant BOLD responses (p<0.001).

Fig. 2 displays the responses to L5b optogenetic stimulation. 1Hz recruited contralateral visual cortex (VC). Both 1Hz and 10Hz evoked bilateral lateral geniculate nucleus (LGN) and superior colliculus (SC) responses. 10Hz induced negative response in ipsilateral VC, and corresponding BOLD temporal signal exhibited an initial positive response before becoming negative. Fig. 3 shows the activations induced by L6b optogenetic stimulation. 1Hz recruited contralateral VC, while both 1Hz and 10Hz induced robust ipsilateral VC activations. Hippocampal activation was also observed during 10Hz stimulation. Fig. 4 details the comparison between stimulation frequencies, and the comparison between stimulation sites. No evoked responses were observed in naïve animals (data not shown).

1Hz L5b stimulation recruits contralateral VC activation (Fig. 2). The optogenetically evoked responses should first occur in L5b of ipsilateral VC, and could subsequently propagate to contralateral VC via L5b callosal projections⁷, or signals from L5b could propagate though local columnar connections to L3^3,4 and then reach contralateral VC via L3 callosal projections⁷. 1Hz and 10Hz L5b stimulation evoked bilateral LGN and SC responses (Fig. 2). Neuronal signals in L5b could transmit to ipsilateral LGN and SC through corticothalamic and corticocollicular projections⁸, respectively. Contralateral SC could receive inputs from ipsilateral SC and contralateral VC through collicular-commissure and corticocollicular projections⁸, respectively. Signals in contralateral VC and SC could propagate to contralateral LGN via corticothalamic and collucular-thalamic projections^8,9.

Heating could cause pseudo negative response, which usually has an instantaneous decrease upon stimulation and takes <5s to reach baseline after stimulation¹⁰. The evoked BOLD change in ipsilateral VC exhibited an initial positive response before becoming negative (Fig. 2). It also took >20s to reach baseline after stimulation. Hence, the negative response was not solely contributed by heating. Although negative BOLD remains unclear, it might be associated with neuronal inhibition that is gradually recruited during excitation in the ipsilateral VC¹¹.

1Hz L6b stimulation recruited contralateral VC (Fig. 3). Neuronal signal in L6b could propagate to L3 and L5b through local columnar connections^3,5, and then to the contralateral VC via callosal projections⁷. For 1Hz and 10Hz, L6b optogenetic stimulation should first evoke response in a column of L6b in the ipsilateral VC, which subsequently propagates to other columns via horizontal connections⁵ resulting in robust large-scale response (Fig. 3). Hippocampal activations observed (Fig. 3) might be associated to cortical-hippocampal projections¹².

Our results indicate that 1Hz optogenetic stimulation evoked stronger contralateral VC response (Figs. 2-4), which might be associated with intrinsic mechanisms of the cortex that oscillates near 1Hz¹³. Our results also suggest that 10Hz stimulation yielded stronger response (Figs. 2-4) which supported the most prominent intrinsic oscillation in infragranular layers at ~10Hz^1,2. Nevertheless, the duty cycle for 10Hz stimulation was higher, which might also yield stronger response. Fixed duty cycle is warranted in future studies.

Layer and frequency specific optogenetic stimulation recruits distinct widespread and large-scale cortical and subcortical activations. Spatiotemporally varying optogenetic stimulation in combination with fMRI presents unique opportunities in studying underlying mechanisms of widespread and large-scale neural circuits and brain networks.