Optogenetic induction of cortical spreading depression in anesthetized and freely behaving mice

Animals

Two- to six-month-old homozygous male and female transgenic mice that expressed ChR2-eYFP fusion protein in layer 5 cortical neurons were used (Supplementary Figure 1(a); strain 7612 – B6.Cg-Tg(Thy1-COP4/EYFP)18Gfng/J; Thy1/ChR2-YFP;^18,19 the Jackson Laboratory, Bar Harbor, ME, USA). For some control experiments, C57BL/6J mice (Harlan, Venray, The Netherlands) of the same age were used. Mice were kept under standard housing conditions with a 12-h light/dark cycle (lights on from 6 a.m. to 6 p.m.) with free access to water and food. All photostimulation experiments were carried out during the light period between 10 a.m. and 5 p.m. The protocol for experiments in Leiden was approved by the “Leiden University Medical Centre Animal Experiments Committee” (DEC 2014-13193), conformed to recommendations of the European Communities Council Directive (2010/63/EU). The protocol for experiments in Los Angeles was approved by the “University of California Los Angeles Institutional Animal Care and Use Committee” (protocol 2005-#021-31), conformed to the National Institutes of Health “Guide for the Care and Use of Laboratory Animals.” All experiments were carried out in accordance with ARRIVE guidelines. All efforts were made to minimize suffering.

Optogenetic induction of CSD combined with monitoring of cortical electrophysiology and blood flow under anesthesia

Optogenetic induction of CSD combined with OIS imaging under anesthesia

Optogenetic induction of CSD combined with monitoring of cortical electrophysiology and blood flow in freely behaving mice

LFP spectrum and vigilance state analysis

For off-line analysis of the cortical LFP power spectrum and vigilance state in relation to CSD, LFP signals from V1 and M1 cortex were low-pass filtered at 100 Hz and down-sampled to 200 Hz (Spike2 v7; CED). Power spectra were computed using a Fast Fourier transform (FFT) routine (Hanning window, 0.25 Hz resolution) per 4-s epoch of the recording starting 20 min before CSD induction. Epochs containing artifacts were excluded from further analysis. To assess frequency band-specific effects of CSD, the average power was calculated per 1-min time bin for five defined spectral bands: delta (δ, 0.75–5 Hz), theta (θ, 5–10 Hz), alpha (α, 10–15 Hz), beta (β, 15–30 Hz), and gamma (γ, 30–45 Hz). In addition, we determined vigilance states (waking, REM sleep, non-REM sleep) by applying a semi-automatic algorithm per 4-s epoch on the basis of the scored behavioral activity and spectral composition of the V1 LFP signal. ²¹ In the algorithm, epochs with active behavior were classified as active waking. The first 10 epochs following cessation of active behavior were classified as passive waking. Remaining epochs were classified as sleeping and further subdivided into REM or non-REM sleep on the basis of the V1 EEG power spectrum: epochs with strong delta activity (>2 × average delta power) were classified as non-REM sleep; high theta (> average theta power) and low delta (< average delta power) activity epochs were classified as REM sleep; sleep epochs not fulfilling the above criteria were unclassified.

Behavioral analyses

Behavior of mice was recorded throughout experiments using a CCD camera (acA1300-60gmNIR, 30 frames/s image acquisition; Basler, Ahrensburg, Germany). Behavior related to the first CSD was analyzed offline using a software-based scoring system (the Observer 12.5; Noldus Information Technology, Wageningen, The Netherlands) for 60 min of recorded video starting 20 min before the CSD-inducing light pulse. Commonly observed rodent behaviors ²² that were analyzed were active (e.g. walking, grooming, eating, scratching, and being attentive to the environment) and inactive (e.g. sitting motionless, lying on the cage floor or sleeping) behaviors. After CSD, additional possible behaviors that were checked for included slow sideway scanning (“orienting” ²³ ) head movements and “wet dog shakes.” ²⁴ Start and end times of observed behavior were scored with 1-s precision allowing for summation of their duration in 1-min time bins. Total active behavior was calculated for each 1-min in bin from the summed duration of all types of behavior except inactivity.

Motor function tests

Mice were operated as described above for recording DC-potential, LFP, and MUA in freely behaving mice. In these animals, motor function was assessed using a wire grip test ²⁵ with modifications to make the test repeatable and specific for assessing forepaw function in combination with V1 photostimulation and electrophysiological recordings (Supplementary Figure 4).

Statistical analysis

Data visualization and statistical testing were performed using GraphPad Prism (GraphPad Software, La Jolla, CA, USA) and Igor Pro v.6.2 (Wavemetrics, Lake Oswego, OR, USA). Data are presented as individual values and mean, except for CSD threshold which is expressed as individual values and median and averaged time series which are expressed as mean ± SEM (CBF, frequency band power, and total active behavior) or mean ± SD (wire grip test). Data were analyzed using paired or unpaired t-tests, Mann–Whitney U test, Wilcoxon signed rank test or repeated measures one-way ANOVA, where appropriate. For the analysis of the wire grip test results, first, grip measurements were averaged from 60 (i.e. the time point at which photostimulation was performed) to 360 s, per forepaw (left/right), per mouse (n = 6), and per condition (subthreshold/suprathreshold). Next, we fitted a linear regression to the resulting 2 × 6 × 2 = 24 averaged values. The independent variables were: mouse identifier, forepaw, condition, and the interaction between condition and forepaw. P < 0.05 was considered significant.

CSD induced by optogenetic stimulation and recorded by intracortical DC-potentials in anesthetized Thy1/ChR2-YFP mice

To assess the possibility of non-invasive induction of CSD by transcranial illumination, we combined optogenetic photostimulation via skull-attached optical fibers with intracortical DC-potential and non-invasive LDF recording (method A₁, Figure 1(a), left panel) in Thy1/ChR2-YFP mice kept under isoflurane anesthesia. Throughout the anesthesia, blood pressure (continuously monitored) and blood gas values were within physiological range (Supplementary Table 1). CSD was successfully induced in V1 cortex using single 4-mW 460-nm (blue) light pulse illumination as evidenced by the characteristic propagating DC-potential shift from V1 to M1 cortex (Figure 1(b)). CSD amplitude (20.5 ± 1.8 mV), duration (59.3 ± 6.3 s), and propagation rate (4.9 ± 0.5 mm/min) (n = 6; Figure 1(c)) were consistent with those induced by other methods, ²⁶ indicating the feasibility of using optogenetics to induce CSD. OPEN IN VIEWER

Optogenetic induction of CSD combined with non-invasive LDF recording in anesthetized Thy1/ChR2-YFP mice

To assess whether light-induced CSD can be reliably detected using non-invasive transcranial LDF, we compared DC-potential and LDF recordings in the same animal (method A₁, Figure 1(a), left panel). Propagation rates measured with DC-potential (4.9 ± 0.5 mm/min) and LDF recordings (5.0 ± 0.5 mm/min) were comparable within animals (n = 6; Figure 1(d)). In a separate group of animals, light-induced CSD was solely recorded with non-invasive LDF recordings (method A₂, Figure 1(a), right panel) to assess whether preparation of burr holes and/or electrode placement affected subsequent CSD characteristics. CSD propagation rates assessed by LDF alone in male mice with intact skull (method A₂) were similar to those assessed by LDF combined with intracortical DC electrodes (method A₁), i.e. 4.8 ± 0.3 versus 5.0 ± 0.5 mm/min, respectively (n = 6, Figure 1(e)). CSD thresholds at 4-mW light intensity were similar whether an invasive intracortical electrode was present (method A₁: 2 to 8 s, or not (method A₂: 3 to 8 s) (both n = 6, Figure 1(f)). Thresholds in male mice were comparable with those in females (method A₂: 3 to 5 seconds; n = 6; P = 0.19, Mann–Whitney U test), so for subsequent experiments data from male and female mice were pooled. Control experiments in Thy1/ChR2-YFP mice showed that CSD was not induced by pulses of red (630 nm) light that do not activate ChR2 or by light-induced rises in cortical temperature (Supplementary Figure 3(a) to (c)).

Characteristics of CSD change when repeated after 20 min

To determine whether optogenetic CSD induction can reliably be repeated under anesthesia, using method A₂, a second CSD threshold at 4-mW light intensity was determined starting 20 min after the first CSD. In all animals, a second CSD could be induced, but this consistently required a longer pulse duration than for the first CSD (n = 12; Figure 1(g)), while the propagation rate was reduced (n = 12; Figure 1(h)). Next, we assessed whether the higher threshold and lower propagation rate for the second CSD were correlated with changes in blood flow. We found that a first light-induced CSD resulted in a characteristic triphasic change in CBF (Figure 1(i)). After an initial hypoperfusion to 52 ± 5% of baseline in the V1 cortex and 43 ± 4% in the M1 cortex, CBF transiently recovered to 94 ± 9% of baseline in the V1 cortex and 88 ± 4% in the M1 cortex followed by a post-CSD oligemia of 53 ± 8% of baseline in the V1 cortex and 44 ± 3% in the M1 cortex. When a second CSD was initiated on top of the preceding CBF response, i.e. during the oligemia phase 20 to 60 min (depending on the actual threshold) after the first CSD, this caused a transient reperfusion, after which CBF returned to a comparable level of oligemia as before the second CSD.

Light-induced CSD under anesthesia displays characteristic vascular responses as determined by OIS imaging

Next we used OIS imaging and electrophysiological recordings to detect vascular and parenchymal changes during light-induced CSD in Thy1/ChR2-YFP mice (n = 3; Figure 2(a) to (c)). In two out of three animals, OIS was performed simultaneously with electrophysiological recordings to confirm that the propagated change in cortical reflectance was accompanied by a deflection in DC-potential (Figure 2(d)). Using 5-mW 470-nm (blue) light pulses, CSD was induced in under 10 s for all five CSDs. A thinned cortical window was used to track changes in vessels, but CSD could also be tracked optically through the intact skull (data not shown). The electrophysiological amplitude of CSD (14.8 ± 1.9 mV) and duration (57.0 ± 8.8 s) (Figure 2(d)) were similar to those previously reported for KCl stimulation. ²⁶ The propagation rate was 3.4 ± 0.5 mm/min (five CSDs), using the filtered image stack to track the CSD wave front over time from individual ROIs (Figure 2(c) and (d)). Kymographs or space-time plots of large and small vessels show vessel contractions that accompany the CSD wave (Figure 2(c)). Larger arterioles or venules showed little or no constriction (e.g. see ROI 6), while smaller vessels showed larger constrictions (range: 7–53%; mean reduction of 32 ± 18%, e.g. see ROI 5). OPEN IN VIEWER

CSD can be induced by light stimulation in freely behaving Thy1/ChR2-YFP mice

To investigate effects of anesthesia on CSD, and to explore the possibility to assess CSD effects on behavior, we induced CSD by photostimulation in freely behaving Thy1/ChR2-YFP mice. To this end, a skull-attached optic fiber above the V1 cortex was combined with intracortical electrodes at V1 and M1 locations for simultaneous detection of DC-potential and neuronal MUA changes (Figure 3(a)). A second optic fiber was attached to the skull overlaying M1 cortex to explore the possibility for non-invasive LDF detection of CSD under freely behaving conditions. Using 460-nm (blue) light pulses, CSD was successfully induced, as evidenced by the characteristic propagating wave of DC-potential deflection (Figure 3(b)). Light-induced CSD in 10 freely behaving male mice showed a CSD amplitude of 17.2 ± 1.9 mV (comparable to CSD under anesthesia; P = 0.27, t-test), a duration of 31.5 ± 3.8 s (shorter than under anesthesia; P = 0.004, t-test) and a propagation rate of 4.89 ± 0.17 mm/min (comparable to CSD under anesthesia; P = 0.99, t-test; Figure 3(c)). In six animals, we determined the CSD induction threshold and found a range of 4 to 9 s with a median of 5 s, illustrating a comparable CSD threshold as for the anesthesia experiments (P = 0.07, Mann–Whitney U test). We also assessed the relation between light pulse duration required for CSD initiation at intensities other than 4 mW and observed an apparent requirement of a minimum pulse duration of 1.5 s and minimum light intensity of 1.5 mW (n = 5, Figure 3(d)). Analysis of MUA and CBF showed that during CSD, the characteristic DC-potential shift coincided with a peak in neuronal MUA that was followed by prolonged suppression of MUA (Figure 4(a)). CBF followed a similar triphasic pattern as observed for the anesthesia experiments. We observed that the post-CSD oligemia coincided with a prolonged secondary negative DC-shift. Induced neuronal MUA during photostimulation of V1 was only detected by nearby V1 (and not M1) electrode pair (insets Figure 4(a)). Control experiments showed that a 5-s pulse of 630-nm (red) light (4 mW or higher) in a Thy1/ChR2-YFP mouse or a 30-s pulse of 460-nm (blue) light (4 mW) in a C57BL/6J mouse did not affect MUA or trigger CSD (Supplementary Figure 3(d) and (e)). OPEN IN VIEWER

OPEN IN VIEWER

Figure 4.

Relationship between light-induced neuronal MUA and CSD in freely behaving Thy1/ChR2-YFP mice. (a) Combined DC-potential and MUA recordings in the visual (V1) and motor (M1) cortex with S1 LDF during V1 460-nm (blue) light stimulation and CSD. Arrival of the CSD wave front at the intracortical electrodes is characterized by a negative deflection of DC-potential coinciding with a peak of intense neuronal firing (i.e. increased MUA) that is followed by neuronal silencing. CBF in freely behaving mice shows a similar triphasic time course as seen under anesthesia (cf. Figure 1(i)). Following the initial CSD wave, the DC-potential trace shows a prolonged secondary negative DC-shift coinciding with prolonged suppression of MUA and post-CSD oligemia. Inset: Details of MUA, DC-potential, and CBF during the subthreshold light pulse in A, illustrating that photostimulation-evoked neuronal MUA is only observed in the illuminated V1 cortical area, and not in the M1 cortex. (b) In four animals in which a second threshold measurement was performed starting 20 min after the first CSD, a comparable threshold value was found for the second compared to the first CSD (n = 4, Wilcoxon signed rank test). (c) Propagation rate for the second CSD was reduced in all animals compared with the first CSD (n = 4, paired t-test). (d) V1 MUA response during the first photostimulation pulse of the second threshold measurement, calculated as % of MUA response during the last subthreshold pulse of the first threshold measurement. (e) Stable CSD threshold during seven repetitive daily CSD threshold assessments. In five freely behaving Thy1/ChR2-YFP mice, a daily threshold measurement was performed for seven consecutive days. CSD threshold changes for consecutive days never exceeded a single step in pulse duration (steps indicated by horizontal lines) and did not show an overt trend between days one and seven.

Repeated CSD during recovery phase has higher threshold and lower MUA

Threshold repeatability was assessed by performing a second CSD threshold measurement 20 min after the first CSD. The second threshold value was higher than the first for all four animals, although this difference was not significant (n = 4; Figure 4(b)). In line with the repeated threshold measurements under anesthesia, the propagation rate was reduced (n = 4; Figure 4(c)). To investigate whether the increased threshold for a second CSD within 60 min after the first CSD is caused by a reduced efficacy of photostimulation, we quantified the neuronal MUA response to photostimulation during the first and second CSD threshold measurements. We observed that levels of MUA during photostimulation pulses of the second threshold were reduced by 27% ± 9 % compared with MUA pulses of the first threshold (n = 4, P = 0.04, paired t-test; Figure 4(d)), indicating a lower cortical neuronal network response to photostimulation. Next, we asked the question whether our optogenetic approach allows for reliable repeated CSD induction over multiple days in freely behaving Thy1/ChR2-YFP mice. We repeated the 4-mW threshold protocol on seven consecutive days and observed stable CSD threshold values with only minor day-to-day variability in required light pulse duration to induce CSD (Figure 4(e)).

Effects of light-induced CSD on vigilance state and behavior

We assessed whether the animal’s vigilance state (as determined from LFP spectral composition and locomotor activity) was influenced by CSD for the first CSD (n = 8). A peak in locomotor activity was observed between 1 and 3 min following CSD induction (Figure 5(a) and (b)). When the CSD wave reached the M1 recording electrode, the LFP spectrum revealed an immediate reduction in all frequency bands that remained low for ∼40 min, except for the delta band that showed fast recovery, coinciding with an enhancement in non-REM sleep. The reduction in LFP power coincided with reduced MUA (Figure 5(a) and (b)). Video-analysis revealed that inactive behaviors (e.g. sitting quietly, lying, sleeping) dominated in the 20-min period preceding CSD (total activity 6.2 ± 2.0 s/min; Figure 5(b), bottom panel). Upon CSD induction, no overt behavioral changes were observed between 0 and 1 min, while between 1 and 3 min all animals displayed a transient peak in behavioral activity (27.2 ± 6.0 s/min) compared to the 20-min pre-CSD period. Detailed video-analysis (example shown in Figure 6(a)) revealed that the peak in active behavior started between 62 and 92 (78 ± 12) s following CSD induction with animals becoming attentive to the environment, or displaying active behaviors such as walking, crawling side- or backwards, displaying turning, or slow scanning movements of the head. These active behaviors could be interrupted by brief periods of immobility. Although observed behaviors following CSD varied, head turning or scanning were observed in five out of eight mice (data not shown). This transient increase in activity ended between 192 and 536 s following CSD induction, when active behavior reduced to a low level of total activity of 1.9 ± 0.6 s/min until the end of the observation period 40 min after CSD. OPEN IN VIEWER

OPEN IN VIEWER

Figure 6.

Effects of light-induced CSD on behavioral activity and cortical function. (a) Still of video (top) and matching electrophysiological V1 and M1 recordings (middle) and ethogram (bottom) showing a representative example of behavior observed before, during, and after CSD induction (see Results section for details). (b) Wire grip test showing equal grip duration for left and right forepaw before and after a subthreshold light pulse over the right V1 cortex, which does not induce CSD (top; P = 0.81, linear regression fit; see Supplementary Figure 4 for test details). Following suprathreshold photostimulation, right forepaw grip duration was not altered (bottom; P = 0.29). In contrast, grip duration of the left (contralateral to the CSD) forepaw was reduced following suprathreshold photostimulation (P < 0.001). The transient reduction is apparent in the data between 30 and 310 s after the suprathreshold photostimulation over the right V1 cortex, the start of which roughly corresponds with the CSD wave reaching the M1 electrode (gray shading indicates the time range in which the CSD wave arrived at the M1 electrode).

Effect of CSD induction on motor function

The start of increased behavioral activity (between 62 and 92 s) following CSD induction in V1 roughly corresponded with the time-point at which CSD was measured at the M1 electrode (between 52 and 67 s). No overt motor dysfunctions were apparent following the short period of increased activity. Separate experiments were therefore performed to assess whether a CSD wave passing the M1 cortex may disrupt motor function. To this end, a repeated forepaw wire grip test was performed (n = 6; Supplementary Figure 4) to assess the ability of a mouse to hold on to a wire with its forepaws before, during, and following optogenetic CSD induction in the right V1 cortex using a suprathreshold 460-nm light pulse. In control experiments, a subthreshold light pulse was used, which did not affect forepaw grip duration (Figure 6(b), top panel). In contrast, loss of left (contralateral to the CSD) forepaw function was evident in the time period after CSD induction, with the start roughly coinciding with the detection of the CSD wave at the M1 electrode (Figure 6(b), bottom panel). Analysis of the data revealed no difference between right and left forepaw grip duration in the subthreshold condition (P = 0.81) and no effect on right forepaw grip duration in the suprathreshold condition (P = 0.29). However, the effect of the suprathreshold condition differed greatly with respect to grip duration between the right and left forepaw (P < 0.001), indicating transient impairment of left forepaw motor function following CSD induction.

Characteristics of non-invasive light-induced CSD are similar to those with invasive induction paradigms

The main electrophysiological characteristics (i.e., amplitude, duration and propagation rate) of light-induced CSD in our optogenetic model, recorded by intracortical electrodes, are consistent with those observed for traditional, invasive induction methods, e.g. KCl-induction^7,26,27 and electrical stimulation^7,28 in mice and rats.^7,29 Comparison of propagation rates of light-induced CSD when recorded by invasive intracortical DC-potential or non-invasive LDF recordings showed that both methods gave comparable results and the presence of burr holes for electrode placement did not affect CSD threshold or propagation rate. In addition, CBF dynamics of light-induced CSD recorded by LDF showed an initial hypoperfusion followed by a short recovery and longer-lasting post-CSD oligemia, consistent with mouse CBF dynamics seen for CSD induced with KCl or electrical stimulation. ²⁶ Additional OIS imaging allowed us to establish that parenchymal and vessel diameter changes during light-induced CSD were also in line with previous studies using KCl or electrically induced CSD, ⁷ and provided spatial and temporal characteristics not obtained by our LDF measurements.

Advantage of non-invasive CSD induction over traditional invasive induction methods

Intracortical KCl injection, ⁷ placement of electrical stimulation electrodes, ²⁸ or the exposure of the cortex for application of KCl-soaked cotton balls ³⁰ have the potential to influence CSD characteristics, because these interventions combined with the need to drill a burr hole at the induction site can evoke CSD by mechanical damage to the cortex. In addition, both supra-physiological concentrations of KCl ³¹ and electrical stimulation ³² can cause considerable local tissue damage, and likely more so in cases when CSD is repeatedly triggered. Our data in anesthetized mice indicate that effects of invasive surgery do not appear to influence threshold and propagation rate of CSD. Invasive procedures may, however, alter effects of CSD on brain function, and behaviors that are relevant to neurological disorders.^2,3 In this study, we used invasive intracortical electrodes to verify the characteristic electrophysiological features of CSD. We also showed that CSD could easily be monitored non-invasively through the intact skull with LDF or OIS approaches, showing their potential for application under awake (head-fixed) conditions.

Advantage of non-invasive CSD induction in freely behaving mice

A non-invasive CSD induction approach offers novel opportunities to investigate the effects of repetitive CSDs in freely behaving animals. Previous studies of CSD in freely behaving rodents have used either single or repetitive injections of KCl through a burr hole.^33,34 The optogenetic model avoids some significant potential artifacts associated with these previous methods. We found that light-induced CSD in the optogenetic model had characteristics comparable to those observed under anesthesia, except for a shorter CSD duration, possibly due to the absence of anesthesia. Given the obvious effects of general anesthesia on brain cellular function and physiology, including reduced neuronal network activity ³⁵ and reduced neurovascular coupling, ³⁶ it is important to be able to examine CSD in the absence of anesthesia. A second advantage of our optogenetic paradigm is the ability to reliably induce CSD on multiple days in freely behaving animals with a consistent threshold for CSD induction for at least seven consecutive days. The model thus can be used to study the behavioral effects (like the shown transient motor dysfunction) of potentially exacerbating or therapeutic interventions over extended periods of time.

Neuronal depolarization suffices to initiate CSD

As optogenetics allows for cell-selective stimulation, our approach can be used to assess the role of individual cell types in the initiation of spreading depression. Although excessive neuronal depolarization has long been assumed to initiate CSD,^3,5 previous studies on neuronal mechanisms of CSD initiation, as discussed above, may have been confounded by tissue damage. The ability to stimulate CSD in Thy1/ChR2-YFP mice, which exclusively express blue light-sensitive ChR2 cation channels ³⁷ in deep layer cortical neurons^18,19 demonstrates that light-driven depolarization of this specific neuronal population can evoke CSD. Earlier studies suggested an increase in extracellular K⁺ above a critical value as the key initiating event that sets off a cascade of neuronal and glial ion transients that may progress into CSD in an all-or-none manner.^3,5 Extracellular K⁺ levels are influenced by various factors, including neuronal activity levels, vascular supply of oxygen and glucose, and astrocytic K⁺ buffering, each of which has been implicated in the initiation of SD.^2,38–40 Although a recent study showed that neuronal activation via sensory afferents can trigger SD when the somatosensory cortex becomes hypoxic during experimental stroke, ⁴¹ our present results provide direct evidence that neuronal depolarization is sufficient for acting as the primary event initiating SD in intact, unchallenged brain tissue. Our data do not exclude the possibility that secondary activation of surrounding neurons (that do not express channelrhodopsin) or astrocytes could also be involved in CSD initiation.

CSD causes long-lasting changes in cortical electrophysiology and blood flow that influence subsequent CSD characteristics

Our light-induced CSD recordings replicated several of the after-effects of CSD that were previously reported for invasively triggered CSD, including the long-lasting post-CSD oligemia.^26,39 Perhaps surprisingly, despite the abovementioned effects of anesthetic agents like isoflurane on cortical neuronal activity and CBF, we found similar values for CSD threshold, amplitude, and propagation rate for anesthetized and freely behaving mice. Notably, for both conditions the threshold for a second CSD (assessed 20 min after the first) was increased, whereas propagation rate was reduced. The changed CSD parameters seem correlated with the observed reduced network activity (evidenced by the reduced MUA during recovery period) and oligemia during post-CSD recovery.

Effects of light-induced CSD on behavior in freely behaving mice

Our non-invasive CSD induction paradigm presents a unique opportunity to investigate changes in behavior related to CSD in non-anesthetized and unrestrained mice. Between 1 and 3 min after the initiation of CSD, a transient increase in active behavior was observed, possibly reflecting transient activation of motor areas given the simultaneous detection of CSD by the M1 electrode. The behaviors during this active period alternated with one or more short periods of immobility, which may reflect freezing-like behavior described upon CSD induction in rats.^24,33 This cannot be confirmed for our experiments since our video analysis did not consistently allow assessment of fixed staring. It is unclear whether the occasional observed head movements and crawling behavior may reflect painful sensation, also because “wet dog shakes” and grooming, which have been suggested to be signs of discomfort or pain, ²⁴ were not observed. In line with the reduced behavioral activity starting 4 to 5 min after CSD induction, our vigilance state analysis showed that despite a general reduction across LFP spectral bands, the delta band showed fast recovery, perhaps even overshoot, that seems to coincide with an enhancement in non-REM sleep, which is in line with elevated delta power with increased sleep reported upon CSD in rats. ⁴² Optogenetically induced CSD in a defined cortical region was shown to disrupt specific behavior as assessed by wire grip tests that revealed clear transient contralateral forepaw wire grabbing impairment coinciding roughly with the CSD wave reaching the M1 cortex.

Future perspectives for optogenetically induced CSD

Our optogenetic CSD model has significant advantages as CSD events can be elicited over multiple days in a non-invasive fashion and can be used to investigate effects of CSD on cortical function, for example by combination with specific behavioral tasks. This provides opportunities for investigating effects of CSD in the context of normal as well as diseased cortical (dys)function. Future applications may lay in migraine and ischemia models, where optogenetically induced spreading depression (or depolarization) may provide clinically relevant insight in aura mechanisms and infarct evolution.^1–3 In addition, our optogenetic model allows for characterization of effects of gene mutations, the influence of trigger factors, and therapeutic interventions relevant to CSD. Optogenetics have been successfully used in neuroscience studies, ¹⁵ ranging from (single-cell) neurophysiology^19,43, to behavior, ⁴⁴ which seems to suggest the introduction of channelrhodopsins does not have serious confounding effect on neuronal function. Finally, the possibility of optogenetics to selectively stimulate (or inhibit) other cell types ¹⁵ is expected to further our understanding of CSD mechanisms.