Synopsis

Keywords: Brain Connectivity, fMRI, preclinical, functional connectivity, neural oscillations, DMN

The default mode network (DMN) is a distributed functional system of the human brain widely studied with fMRI due to its involvement in advanced cognitive processes and its dysregulation in a variety of brain disorders. Causal perturbations of this network in physiologically accessible species are critically required to probe its circuit organization and the underpinnings of its (dys)function. Here we show that tapered fiber optogenetic technology enables the reliable stimulation of key DMN nodes in a frequency dependent fashion, overcoming the limitations of traditional optogenetic approaches.

Introduction

Resting state fMRI (rsfMRI) is used to explore the intrinsic network organization of the human brain in the absence of explicit tasks. Spontaneous low-frequency oscillations in BOLD fMRI signal exhibit reproducible and anatomically specific patterns of correlations between different brain areas, defining large-scale, anatomically distributed functional connectivity (FC) networks. A widely investigated FC network of the human brain is the default mode network (DMN), a system thought to be a key substrate of internal modes of cognition such as empathy, recollection and imagination, conceptual processing and conscious awareness¹. The robust evidence for altered DMN connectivity under pathological states¹ has spurred research into the presence of anatomically conserved DMN precursors in accessible species. Recent years has witnessed the discovery of an analogous rodent DMN system involving evolutionarily conserved associative prefrontal and peri-hippocampal cortices². Cell-type specific neural perturbation of the DMN in rodents offers the opportunity to investigate the core constituents and neurobiological underpinnings of DMN (dys)function. However, the widely distributed topography and associative nature of the DMN pose challenges for opotgenetic stimulation. Traditional flat fiber technology is typically limited to focal manipulation of individual circuit components due to the risk of tissue heating and spurious hemodynamic responses³. To attain homogeneous photostimulation of large volumes and network-scale manipulations of the DMN, we propose the combined use of mouse fMRI and tapered fiber (TF) technology⁴. Here, we describe and thoroughly validate the use of TF technology to attain network level manipulations of the DMN through photostimulation of the mouse medial prefrontal cortex (mPFC), one of its evolutionarily-conserved core constituents⁵.

Methods

For validation of the TF system, Thy1-ChR2 mice (B6.Cg-Tg(Thy1-COP4/EYFP)18Gfng/J; Jax #7612; expressing channelrhodopsin-2 [ChR2] under the Thy1 promoter) were used. Proof-of-concept experiments exploring the impact of different stimulation parameters were instead performed in mice injected with a viral vector expressing ChR2 under the CamkIIα promoter (AAV5-CamkIIα-hChR2(H134R)-EYFP, Addgene #26969). The single TF was implanted in the mPFC at a 15^o angle (AP +1.9 mm; ML +/- 0.6mm; DV –2.1 mm; Figure 1). Blood oxygen level dependent (BOLD) images were acquired on a 7T scanner (Bruker) using an echo planar imaging gradient echo (EPI-GE) sequence with the following parameters: TE=15ms, TR=1s, flip angle=60^o, NR=390, matrix size=98x98, slice number=18, field of view=2.3x2.3x9.9mm. All scans were acquired under anesthetic; a combination of medetomidine (0.1 mg/kg/h i.a) and isoflurane (0.5%) (med-iso) was used for the majority of experiments. Photostimulation was performed by pulsing light either in a block design (15ms pulses at a frequency of 20Hz for 10s, every 60s) or continuously. BOLD images were reconstructed and preprocessed as previously described⁶. The evoked response from block design stimulation was characterized with a GLMM⁷, and by modelling the hemodynamic response function as a Fourier basis set convolved with a boxcar function representing the stimulation paradigm. FC differences due to continuous stimulation were evaluated compared to opsin free controls.

Results and discussion

Single probe dual hemisphere illumination was achieved by implanting a single TF into the mPFC at a 15^o angle, crossing the midline such that light was delivered on both sides. The success of this method was demonstrated in slice by light emission induced fluorescence from a tapered fiber implanted into a coronal section from a Thy-ChR2 mouse (Fig. 1). In keeping with this, in vivo optogenetic-fMRI studies revealed that stimulation of pyramidal cells with a single TF in the mPFC under either promoter (Thy1: Fig. 2; CamkIIα: not shown) resulted in exquisitely bilateral stimulation of key DMN afferents. The largest response was detected in the thalamus, a region recently recognized as an important component of the DMN. Notably, the bilateral response achieved with a single low-invasive TF was comparable to that obtained with a canonical dual flat fiber configuration, and distinct from the unilateral response obtained with a single flat fiber implant (Fig. 3). We also tested for possible heat or visual induced responses by implanting control mice (without ChR2) with a TF. Contrary to previous observations with flat fibers^3,8, we found that TFs could be used to perform widespread illumination without any observable changes in BOLD response (Fig. 2, green/dashed line). Similarly corroborating the specificity of the mapped effects, blood pressure recordings obtained during stimulation under different anesthetics (med-iso vs. halothane) showed that the evoked response was uncoupled from peripheral cardiovascular changes. Finally, proof-of-concept experiments were performed using this technology, showing that continuous rhythmic stimulation of the mPFC resulted in changes to the FC of cortical and subcortical DMN substrates that were dependent on the stimulation frequency. Importantly, while the evoked response from block stimulation could be predicted from mPFC projections, continuous stimulation engaged components of the DMN above and beyond what would be expected from the structural connectome⁹.

Conclusions

We show that it is possible to use a single tapered fiber to reliably photostimulate the mouse mPFC and bilaterally engage synchronous activity within anatomical nodes of the DMN as predicted by the distributed wiring diagram of this region⁵. Importantly, proof-of-concept studies showed that anatomical substrates engaged by continuous or rhythmic mPFC stimulation are anatomically and functionally different, hence paving the way to the investigation of the dynamic rules governing the organization of this translationally relevant network.

Acknowledgements

This work was supported by the European Research Council (ERC – DISCONN; no. 802371 to A.G.)