Avalilação PREreview de Context-dependent presynaptic inhibition of somatostatin interneuron inputs to Layer 1 of the visual cortex

In this preprint Alyahyay et al. describe a novel inhibitory circuit motif in layer 1 of the mouse visual cortex, in which NDNF interneurons presynaptically inhibit the axonal outputs of SST interneurons during locomotion and low-contrast visual conditions. Although regarded as a potential modulator of other cortical interneurons like PV and VIP interneurons, NDNF cell interactions with SST interneurons have yet to be systematically studied. Using a combination of physiological, histological and computational approaches, the authors report that NDNF interneurons mostly form axo-axonic, rather than axo-somatic, synapses with SST neurons. Moreover, functional investigation of NDNF-SST connectivity using slice-electrophysiology revealed that this motif participates in the disinhibition of the pyramidal targets of SST neurons. Finally, the authors show that the engagement of this NDNF-SST motif is dynamically governed by behavioral context, with NDNF interneurons inhibiting SST interneurons most strongly during low-contrast visual stimulation and locomotion.

This manuscript addressed a timely topic in neuroscience regarding how complex local network of cortical interneurons modulate context-dependent high-order processing. The wide variety of experimental approaches provided strong and converging pieces of evidence to support the conclusions drawn. The proposed novel NDNF-SST motif contributed to a more holistic understanding of the neuronal circuitry in cortex and their function. However, the conclusion of the study could be fortified further by addressing the following concerns.

Major concerns:

1). The authors concluded that this NDNF-SST motif is composed of direct NDNF-SST synaptic communication. However, given the limitations of in vivo experiments that require visual stimulation, the possibility that a feedforward inhibition through NDNF-pyramidal-SST could also affect the SST neurons is not completely controlled.  Could the authors provide discussion regarding the existence of this feedforward inhibition, and whether recording evoked inhibitory responses directly from the soma / axon of the SST neurons in response to opto-stimulating NDNF neurons has been attempted to validate the direct synaptic transmission.

2). In the figure legends and methods section for the calcium imaging data, the authors state both animal number as N and ROI number as n. It is not made clear within the figure legends whether animal numbers or ROIs were utilized as biological replicates in running the statistics. Since ROIs are not a meaningful biological unit, my understanding is that these represent technical replicates while the animals represent biological replicates. This could be clarified to avoid concerns related to pseudoreplication, in which technical replicates rather than biological replicates are used for statistical analysis.

Minor points:

1). Page 2, Line 13-15: “Studies using whole cell patch clamp recordings indicate that SST cINs strongly connect to NDNF cINs, while the reciprocal connection from NDNF cINs to SST somas is nearly absent”. Could the authors be more specific regarding connection, i.e. does this mean that SST neurons synaptically project to NDNF neurons? Or through feedforward mechanism?

2). The clarity of under what condition is NDNF-SST motif engaged could be greatly enhanced by a summary figure comparing the SST activity with all four combinations (low contrast +stationary, high contrast + stationary, low contrast + locomotion, high contrast + locomotion) with proper statistical methods (ANOVA etc.).

3). Figure 3F,G,I and J. These panels are important for the clarification of the heterogeneity of the pyramidal responses under different Ca+ conditions. However, these panels could add more weight to support the notion of disinhibition by adding data points from oIPSC or C-oIPSC into the same panels, which would provide more information about the context-dependency of these NDNF-SST projections when the output of SST neurons vary.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.