Discussion
The main conclusion of this study is that a significant number of discontinuous light stimulation patterns, commonly used in optogenetic experiments, can affect neuronal electrophysiological activity. However, their impact is, in general, smaller than that produced by continuous light stimulation. It also highlights how the different parameters of the stimulation impact the light action, providing several guidelines to minimize the artefactual effects associated with the use of light in neuroscientific investigations. Indeed, a bibliographic analysis done on the first 50 discontinuous patterns retrieved when using the query “Optogenetics” on PubMed showed that the light parameters investigated here are among the most commonly used in optogenetic experiments (https://osf.io/anvsj). We used the change in AP amplitude and latency and the generation of an outward current as a read-out for light action, all of which were modified by continuous light stimulation (Ait Ouareset al. , 2019). We found that patterned light consistently reduced AP amplitude without significantly altering AP latency in all cell types and produced an outward current only in MCs, PYRs and GCs. The lack of effect on AP latency is likely related to the low statistical power of the pre-registered analytical method, since exploratory analysis shown in supplementary table 3 suggest that AP latency does indeed slightly increase during patterned light stimulation. The effects of light on membrane current and AP amplitude are small, with a change of only a few pA for membrane current and a ~0.5 mV reduction in AP amplitude for the pattern with the highest effect. However, even small variations in AP amplitude can consistently modify synaptic transmission (Rama et al. , 2015). Moreover, discontinuous light produced a strong reduction in firing frequency for a subpopulation of MCs when neurons where depolarized to a membrane potential that allow spontaneous activity. A possible reason for the heterogeneity of this effect might be explained by a diversity of biophysical properties among different MCs (Padmanabhan & Urban, 2010; Fourcaud-Trocmé et al. , 2022). However, the effects of discontinuous light where in general lower than those produced by continuous stimulation showing that patterned stimulation can reduce the artifactual effect of light. Both the induction of the outward current and the reduction in AP amplitude induced by discontinuous light occur rapidly after few pulses and quickly reach a plateau. Therefore, reducing the number of stimuli does not appear to be an optimal strategy to reduce the artefactual effect of light. On the other hand, the light effect can be attenuated by reducing the stimulation power, the duty cycle and, in some cases, the pulse duration. Note that varying the frequency of the light stimulation does not change the artefactual effects as long as the duty cycle is the same. This is particularly important for experiments whose goal is to induce high frequency firing by optogenetic stimulation. As a rule, we recommend adapting the pulse duration to the stimulation frequency in order to keep the duty cycle at or below 5%. Using a light power of 1 mW or less also appears to be an effective means of preventing unspecific effects during discontinuous stimulation compared to the higher powers, which induce an outward current (see supplementary table 1). The reduction in AP amplitude induced by pattern 4 is observed in all cell types, whereas an outward current is induced only in FSIs and MSNs, suggesting a cell-specific effect of light. The greater reduction in the discharge frequency of FSIs induced by light compared to pyramidal neurons (Owen et al. , 2019) also suggests a differential effect of light across cell types. However, we did not observe such a difference here where all experiments were done in the presence of ionotropic glutamatergic receptor antagonists. Since light appears to reduce glutamatergic synaptic transmission (Ait-Ouares et al., 2019), the pronounced decrease in FSI firing observed by Owen et al., (2019) may be related to a reduction in their excitatory inputs. However, we cannot rule out the possibility that the weaker effect observed here is due to the presence of the tomato fluorescent protein in the recorded FSI. The covariation between the change in tissue temperature and the effect of light suggests a causal relationship between the two factors, as previously reported (Stujenske et al. , 2015; Senova et al. , 2017; Ait Ouares et al. , 2019; Owen et al. , 2019). The light-induced temperature changes observed in this study (between 0.1 and 0.8 °C) remain below the threshold for tissue damage but may modulate a set of voltage-dependent channels involved in neuronal discharge (Podgorski & Ranganathan, 2016; Ait Ouares et al. , 2019). Our results suggest that light is able to modulate different type of K+ channels beside the inwardly rectifying potassium channels (Kir) (Owen et al. , 2019). For instance, the reduction in AP duration suggests an increase in Kv conductance’s responsible for AP repolarization (Bean et al., 2007) while the reduction in AHP amplitude suggests an impact of light on the calcium- and voltage-dependent potassium channels (Sah & Faber, 2002). The light-induced reduction in AHP amplitude after a single AP contrasts with the increase in AHP measured at the end of a long depolarization that we observed previously (Ait Ouares et al. , 2019). This result suggests that light may oppositely modulate the K+ channels involved in the fast-medium and slow AHP. However, because the AHP in MCs is modulated by recurrent glutamatergic transmission (Duménieu et al. , 2015), we cannot rule out that the light-induced reduction in AHP amplitude is partially due to the presence of ionotropic glutamatergic receptor antagonists. Finally, because AP initiation in MCs is promoted by hyperpolarizing events (Fourcaud-Trocmé et al. , 2022), the light-induced decrease in AHP could also contribute to the reduction of the firing frequency of these cells.