Stress is a critical risk factor for developing neuropsychiatric disorders that induce long-lasting physical and psychological changes. Stress also activates the hypothalamus-pituitary-adrenal (HPA) axis leading to enhanced glucocorticoid production [1]. It is well known that a series of exposures to glucocorticoids affects dendritic remodeling and synaptic changes in the brain [2-4], causing functional changes in neural networks [5, 6].

A single acute stressor can cause structural and functional changes in the neural network, including altered glutamate/GABA transmission [7] and an imbalance between glutamatergic and GABAergic neurons in the amygdala [8]; furthermore, a single restraint acute stress can cause changes in excitation-inhibition (E/I) balance of neural responses leading to the reduction of excitatory-inhibitory coherence of neural network [9]. Additionally, acute stress not only alters the vascular and neural temporal dynamics but also differentially affects a regional synaptic transmission, i.e., downregulation of glutamatergic neurons in amygdala under acute glucocorticoid application [8], disruption of glutamate/GABA transmission in somatosensory cortex under acute stress [9] and increase of glutamatergic transmission in the medial prefrontal cortex (mPFC) and hippocampus through enhanced glutamate release [10-12]. These findings indicate that acute stress influences isolated brain regions and modulates interactions across a wide brain network. Therefore, a systematic approach is needed for understanding the functional connectivity of the broader network rather than solely focusing on individual brain regions.

Functional magnetic resonance imaging (fMRI) is a neuroimaging technique that can measure the functional connectivity of the brain network based on blood oxygenation level changes, serving as a surrogate for neural signals. The non-invasive nature of fMRI provides important biomarkers for understanding the state of neurological diseases of animals and humans. One of the hemodynamic imaging methods like fMRI is intrinsic signal optical imaging. In vivo intrinsic optical signal imaging can measure various blood oxygenated states, including total hemoglobin, deoxyhemoglobin, and oxyhemoglobin, in response to changes in neural activity using different wavelength filters [13]. When utilizing thinned skull techniques, intrinsic optical signal imaging can measure hemodynamic signal changes occurring throughout the broad range of brain regions with higher spatio-temporal resolution than fMRI [14]. Thus, combining hemodynamic optical imaging techniques with fluorescent imaging provides a unique opportunity for simultaneously monitoring hemodynamic signals and neural calcium signals from brain regions [15].

In the study of Han and her colleagues (2020), a single acute stressor altered NMDA-evoked vasodynamics of penetrating arterioles as well as the calcium activity of excitatory neurons in the somatosensory cortex. The peak amplitude of both vascular and neural responses was lower in acutely stressed mice than in control mice and the time to peak of responses was delayed in stressed mice. This study also highlights that acute stress can affect not only the HPA axis but also cortical regions in the brain; however, there are currently no reports on how acute stress affects the broader cortical area with concomitant calcium and hemodynamic imaging. Therefore, we utilized dual optical imaging with thinned skull techniques to investigate the effects of acute stress on the brain network level. We aimed to investigate whether acute stress induces temporal profile changes in cortical response within each brain region, leading to connectivity changes during normal sensory processing.

Thinned skull surgery and acute restraint stressed model

Eight-week-old female Thy1-GCaMP6f mice (Tg-Thy1-GCaMP6f-GP5.17DKim/J, JAX 025393, The Jackson Laboratory) were used in this study. The animals were maintained with a 12 h dark/light cycle (light on 9:00 P.M.), 50%~60% humidity, a temperature of 24°C~25°C, and foods in ad libitum. All experimental procedures were approved by the Institutional Animal Care and Use Committee of Sungkyunkwan University (SKKUIACUC2021-12-23-1). Twenty-eight Thy1-GCaMP6f mice were used for optical imaging (control: n=10; stress: n=18) and ten Thy1-GCaMP6f mice were used for ELISA (control: n=6; stress: n=6).

The mice were initially anesthetized with 3% isoflurane in an induction chamber, and anesthesia was maintained at 1%~1.5% isoflurane during surgical procedures. Body temperature was kept at approximately 37°C using a temperature-controlled heating pad. The scalp over the craniotomy area was resected with a surgical blade. Thinned skull surgery was performed on the frontal and parietal lobes of the mice using a dental drill (Microtorque II, Ram Products) with constant saline irrigation to remove skull vessels or opaque skulls as much as possible while avoiding excessive drilling of the sutures. Following the thinning, a thin layer of cyanoacrylate (gel-type Loctite^® Super Glue) was applied to the surface of the skull to prevent bone re-growth and provide mechanical protection to the surgical area. A screw was put on the right olfactory cortex, and a metal holding frame was glued onto the skull and fixed with dental resin to prevent animal head motions during the imaging experiments. Mice were given a week for recovery, during which they received 1 mpk of meloxicam, a non-steroidal analgesic, and 5 mpk of Baytril, an animal antibiotic, injected subcutaneously to manage pain. Acute stress was induced by restraining the mice’s motion for 30 min before optical imaging. The mice were placed into well-ventilated plastic bags (DecapiCones, Braintree Scientific) with sealed exits using a cable tie to immobilize them. During the restraint, the animals were deprived of water and food intake. The control mice were allowed to move freely in their cages. This acute restraint stress protocol induces the alteration of corticosterone levels, leading to various behavioral changes and functional alterations in neurons [10, 14, 15].

Blood sampling and ELISA

To determine the corticosterone level in mice plasma, mice (control: n=6; stress: n=6) were sacrificed immediately after the 30-minute restraint stress under Ketamine-Xylazine anesthesia (100 mg/kg+10 mg/kg, i.p.). Heparin-coated tubes (BD Vacutainer, Becton Dickinson) were used to collect trunk blood samples. The blood samples were centrifuged twice at 5000 rpm for 10 min at 4°C to obtain plasma samples. An ELISA kit (Assaypro) was used to measure the corticosterone concentration in the plasma. The absorbance value was measured using a microplate reader (Synergy HT, BioTek) at a wavelength of 450 nm. The sample concentrations were calculated using a standard curve generated by standard solutions. The plasma corticosterone concentration in blood samples from control and acutely stressed mice was measured at 9:00 am. Corticosterone levels in mice exhibit a distinct circadian pattern within a daytime maximum range [16].

In vivo optical imaging

After a thirty-minute recovery period following restraint stress, mice were anesthetized with Ketamine-Xylazine (I.p.100 mg/kg+10mg/kg, i.p.) for a wide-field optical mapping session [17]. The whiskers on the right side of the face were trimmed except for C2 and C3 whiskers. Imaging was conducted using an LED-based wide-field optical mapping system, as described in [18], utilizing an Andor sCMOS Zyla 5.5 camera with Nikon 60 mm f/ 2.8D AF Micro-NIKKOR lens. Two band-pass filters (460/60, 530/43 Semrock bandpass filter) were used along with a high-power mounted LED to prevent the excitation light from leakage and define a narrow spectral band. A dichroic mirror allowed LEDs to produce a single beam which is positioned to directly illuminate the brain.

The 505 nm long pass filter (DMLP505R-Thorlabs), in front of the camera, enabled the alternated imaging of hemodynamics and GCaMP6f fluorescence. During 460 nm illumination, a fluorescence signal was detected, during 530 nm illumination, the isosbestic point of oxy- (HbO) and deoxy-hemoglobin (HbR), total hemoglobin (HbT) signal was detected. Reflectance (HbT) and fluorescence (Ca²⁺) signals were simultaneously measured using the Master9 pulse stimulator. Images of reflectance and fluorescence were acquired at a frame rate of 10 Hz for each LED wavelength.

Whisker stimulation was given for 4 seconds at 5 Hz (picospritzer III, Parker Hannifin) following a 10-second baseline acquisition time, with a 46-second recovery time after stimulation controlled with a pulse stimulator (Master 9, A.M.P.I., Israel). Each trial lasted for 1 minute and was repeated 15~20 times for thinned-skull mice. Trial-averaged data were used for the imaging analysis. Physiological conditions (i.e., heart rate, pO2, and respiration rate) were monitored using a paw sensor (PhysioSuite, Kent Scientific) to ensure the normal ranges of anesthesia under Ketamine-Xylazine throughout the experiment.

Imaging analysis

The imaging data underwent processing using the modified algorithm from “An open source statistical and data processing toolbox for in mice” [19, 20], along with custom MATLAB code (MathWorks, USA). The processing of Ca²⁺ and HbT data involved four to five steps respectively: Seed ROI setting, application of 2D Gaussian blurring for smoothing, baseline normalization, and boxcar averaging. Regression analysis was performed for calcium data only.

Our 14 seed regions of interest (ROIs), including primary motor (M1), secondary motor cortex (M2), primary somatosensory forelimb (S1FL), secondary somatosensory (S2), retrosplenial cortex (Ret), primary visual cortex (V1), primary somatosensory barrel field (S1BF) of both hemispheres, were selected based on their anatomical distances from bregma and lambda points. Each of these seeds is illustrated in Fig. 1C and has a size of 5×5 pixels. In addition, four surrounding 5×5 ROIs, with their vertex pixel overlapping the center pixel of the seed ROIs, were selected and averaged with the seed ROIs to elevate the confidence level of anatomy-based seed ROIs.

Time series reflectance and fluorescence images underwent processing using Gaussian filters through a custom-written MATLAB code. Then, the images were normalized based on the 10-second baseline acquisition data for each image. The acquired reflectance and fluorescence data were further preprocessed with 2-frame boxcar averaging. Specifically, for calcium fluorescence data, a calcium signal regression was performed on all fluorescence images by dividing the fluorescence ratio by the reflection ratio. This step was performed to eliminate any possible contamination arising from hemodynamic signals. The regression process operated on the assumption that the properties of the tissue (including changes in absorption) are consistent between excitation and emission wavelength. It was assumed that the path length of the 530 nm diffuse reflectance light is approximately equal to the sum of the excitation and emission path length, according to the single-wavelength method [18].

Correlation analysis: Cross-correlation, Pearson correlation, and Spearman correlation coefficient

The custom MATLAB code was used to calculate the maximum cross-correlation coefficient (FC_CC; functional connectivity measured by maximum cross-correlation coefficients) between each set of reflectance and fluorescence data. These values were used to estimate the similarity between three different time-series signals and infer any changes in pathways. The imaging data was divided into three time periods: baseline (0~10s), stimulation (10~14s), and recovery period. Recovery period for calcium was set for 15 to 19s by 1s interval and for HbT was set for 16 to 20s by 2s interval. The durations of these periods were determined based on the average Full Width Half Maximum (FWHM) value at seed #14. Due to a 1-second delay in the FWHM reaching time for HbT compared to Ca²⁺ at seed #14, the analysis for HbT was conducted with a 2-second interval instead of the 1-second interval used for Ca²⁺.

Pearson correlation coefficient (FC_P; functional connectivity measured by Pearson correlation coefficients) was calculated using a custom-made MATLAB code to measure the temporal dependency between neural activation and hemodynamic responses in distinct brain regions [21]. The three-time periods and time-windows were set in the same way of cross-correlation analysis. As recovery phase of calcium signal is included in stimulation period, no additional analysis has been made in the FC_P map analysis for fluorescence data from the recovery phase. Seed #1, #3, #5 are excluded from the region of interest (ROI) at HbT analysis of cross-correlation and Pearson correlation, considering that these seeds contain no onset and peak value.

Spearman correlation [Supp. 1] was used to calculate the coefficient of two variables from stressed and control group’s FC_P values rank, to find possible alteration of FC_P by acute stress.

In compliance with the normality results, Mann–Whitney U tests or independent t tests were applied for comparisons of two groups. To determine the correlation coefficient between two variables, Spearman's coefficient (r) or Pearson's coefficient (r) was calculated according to the normality outcomes. A value of p<0.05 was considered statistically significant for comparisons of two groups. The statistical analysis was conducted with MATLAB.

Neural and hemodynamic changes upon sensory stimulation in acutely stressed group and control group

To induce acute stress, mice were immobilized in plastic bags for 30 min (Fig. 1A) [22]. The level of corticosterone in plasma was measured through ELISA, and the corticosterone concentration was higher in stressed group than in the control group (control vs stress: 77.7 ± 23.2 ng/ml vs 142.8 ± 23.6 ng/ml, Mann-Whitney U test, **p=0.0022, control: n=6; stress: n=6), indicating that the animals were under a stress condition (Fig. 1B). Seven different ROIs were chosen from each hemisphere based on cortical anatomy atlas (Fig. 1C), and from total 14 ROIs, calcium fluorescence (ΔF/F) and hemodynamic signals (ΔR/R) were quantified as described in the method. First, we focused on the left S1BF and S2 regions (seed #14, #11 from Fig. 1C) to monitor the calcium and hemodynamic signals under sensory stimulation (Fig. 1D~G). Sensory whisker stimulation is relayed to the somatosensory barrel field via the thalamus [23]. GCaMP6f calcium signal activation indicates changes in calcium signal from the excitatory neuron. The calcium signal of excitatory neurons did not show significant differentiation between the control group and the acutely stressed group, as evidenced by the quantified onset time, peak values, time-to-peak, and area under the curve (AUC) measurements from each group (Fig. 1H). The hemodynamic signals in the primary and secondary somatosensory cortex were observed to be statistically insignificant between the mice groups (Fig. 1I). However, one clear distinction between the two groups of mice was evident. In the wide field of view, the evoked hemodynamic responses and neural activity exhibit a distinct pattern, depending on whether it is under acute stress or in a control condition. Analyzing the differences between groups revealed variations in temporal patterns. (Fig. 1J, K). Therefore, we analyzed the connectivity between cortical brain areas and quantified the altered brain network induced by acute restraint stress.

Temporal FC changes in acutely stressed group with cross-correlation

Analysis of neural and perfusion signals in the ROIs of S1BF and S2 areas indicates that the projection via the hypothalamus, as presented by neural and hemodynamic responses, remains largely unchanged by acute stress. In contrast to these individual ROIs analyses, we applied cross-correlation analysis to assess the similarity and coherence between distinct cortical areas at different time periods, including stimulation time, peak time, and recovery time. Then, we calculated the time lags with the maximum cross-correlation between two seeding areas at specific time periods and compared them using the Mann-Whitney U test between the control and acutely stressed groups (Fig. 2A, B), repeating this process for all seed pairs and time periods.

A significant time lag in connectivity of the calcium signal from the excitatory neuron was observed between different seed pairs: #11 vs. #13 and #13 vs. #14 during the stimulation period (10~14s) (Fig. 2C). In the control group, seed #13 showed a maximum cross-correlation with time lags ~0.14s faster on average than seed #11 and ~0.22s slower on average than seed #14. But in the stressed group, seed #13 displayed an average delay of ~0.7s compared to seed #11 and 1.483s compared to seed #14. However, there were no significant alterations in FC_CC. On the other hands, differences were observed in the time lags generating maximum cross-correlation for regions that displayed a moderate positive correlation in both the control and stress groups. In the control group, seed #11 and #13 showed minimal differences (all entities except one showed a 0s delay), whereas in the acutely stressed group, seed #13 was 0.7s slower than seed #11 on average. For seed #14 and #13, in the control group, #13 lagged by 0.22s behind #14, but under stress conditions, this delay significantly increased to 1.483s. Considering that the time-to-peak remained unchanged in seed #13, it suggests that the recovery of barrel projection in the same hemisphere, initiated from seed #14, was delayed in the stressed group.

The analysis of the hemodynamic signal did not show any significant alterations in connectivity during the stimulation period from 10~14s. However, differences in time lags with maximum cross-correlation of hemodynamic signal between the ROIs were observed during the peak period from 14~16s and the recovery period from 16~20s (Fig. 2D, E). The relationship between two significantly changed ROIs during the peak period from 14~16s is depicted in Fig. 2D. In the control group, seed #14 exhibited an average delay of about 0.8s compared to seed #10 (FC_CC: 0.353). Conversely, in the stressed group, seed #14 displayed a delay of around 0.2s compared to seed #10 (FC_CC: 0.195). There was no significant difference in the area under the curve (AUC) during this period. Seed #14 exhibited an average time-to-peak value of 15.1s (control) and 15.4s (stress), while seed #10 showed peak values at 13.4s (control) and 14.3s (stress). Seed #10 demonstrated a faster time-to-peak compared to seed #14 only in the control group (p=0.0114, Mann-Whitney U test). Additionally, in this time period, seed #14 in both the control and stressed groups, as well as seed #10 in the stressed group, exhibited peaks. However, the control group's seed #10 displayed recovery during this time period, indicating that seed #10 had a slower time-to-peak and recovery in the stressed group compared to the control group. In the control group, seed #12 was approximately 0.41s faster on average than seed #4 (FC_CC: 0.211), while in the stressed group, seed #12 lagged by around 0.044s compared to seed #4 (FC_CC: 0.083). Given that seed #12 in the stressed group displayed two peaks on average at 12.5s and 17.8s, it suggests a delay in recovery at seed #12 in the stressed group. The time course analysis of neural and hemodynamic signals in S1BF and S2 regions showed minimal changes under acute stress, while cross-correlation analysis revealed altered coherence between cortical areas with diverse time lags. Notably, delays in the connectivity between specific seed pairs were observed in the calcium signaling pathways, while distinct time lags and delays were noted in the hemodynamic signal connectivity during peak and recovery periods under tress condition.

FC analysis with Pearson correlation suggests that acute stress affects brain-wide connectivity

Maximum cross-correlation analysis determined the time lags that revealed the highest similarity by shifting one seed’s timeline relative to others. This analysis specifically illustrates time points that maximize the similarity between two seeds. But, this analysis does not capture the entire time lags. Therefore, Pearson’s correlation was employed to estimate the linear correlation between the time series data of two seeds. By calculating the overall strength of FC_P between seeds during different time periods, including baseline (1~10s), stimulation (10~14s) and recovery (Ca²⁺: 15~19s; HbT: 16~20s), the FC_P maps for Ca²⁺ were derived from average data of calculated value at 1-second intervals (10 frames) (Fig. 3A) and for HbT at 2-second intervals (20 frames) (Fig. 3B).

The FC_P map exhibited changes in Ca²⁺ response between the control and stressed groups in the stimulation period 10~14s (Fig. 3C). A previous cross-correlation analysis indicated a time lag of the signal originating from seed #14 in the stressed group, highlighting a significant functional change in the connection between #11 and #14. In the 10~11s and 13~14s, the correlation coefficients of #11 and #14 are respectively 0.699 and 0.641 in control and 0.66 and 0.517 in the stressed group, indicating highly correlated connections with no significant difference (Mann-Whitney U test, p=0.79, 0.35). Also, at 11~12s time period, they displayed moderate correlation without significant difference (control=0.47 vs. stress=0.49, Mann-Whitney U test, p=0.94). However, in stressed group, a significant functional change occurs in 12~13s, characterized by a relatively decreased FC_P (control=0.83 vs. stress=0.57, Mann-Whitney U test, p=0.015). In addition, two other connectivity pairs #3 & #10 and #3 & #9 were affected by stress conditions during the stimulation period (p=0.037, Mann-Whitney U test). These connections are moderately correlated in the control group (0.45952 & 0.46374) but exhibited decreased correlation values in the stressed group (0.083 & 0.11856). All other connections between seeds in these periods did not show any significant alterations [Supp. 2]. Thus, under acute restraint stress, the cortex demonstrated a decline in FC_P, particularly within the posterior intra-connection of the contralateral side where the barrel signal originates, and in the anterior inter-connection between the hemispheres

No significant FC_P alterations in HbT response were observed during 10~14s (Mann-Whitney U test of time lags with maximum cross-correlation, and onset). However, there was a significant alteration observed in the FC_P map based on Pearson correlation coefficients (p<0.05). These changes were primarily observed in the FC_P between seed pairs; #11 & #7, #8, #10 and #7 & #6, #8, #9, #10, respectively (Fig. 3D). Notably, the stressed group exhibited higher FC_P in the areas of seeds pairs; #7 & #11, #8 & #11, and #10 & #11, while the control group displayed higher connectivity in areas such as seed pairs; #6 & #7, #7 & #8, #7 & #9, and #7 & #10. The contralateral-cortical FC_P is strengthened, while inter-cortical FC_P, except seeds pair #7 & #11 area, is weakened in the stressed group compared to the control group. FC_P in the contralateral-hemisphere between two groups were changed, in seed pairs #10 & #11, #8 & #11, (Mann-Whitney U test; p<0.05: *0.0103, **0.005 respectively), and the stressed group has higher FC_P than the control group. While no significant differences were observed in the onset time and time to peak between the control and stressed group (Fig. 1I), it is the slope of the stimuli response that accounts for the differences in FC_P between the two groups. Additionally, unlike M1 and S1FL, FC_P between seeds #9 & #11 remained unaffected. This result suggested that M2 has weaker FC_P with stimuli point area compared to M1 and S1FL. M2 is located at a higher hierarchical order than M1.

In the inter-cortical FC_P, Mann-Whitney U test in the seed pairs #7 & #8, #7 & #9, #7 & #10 and #7 & #11 showed significant alterations (*p<0.05, **0.008, *0.041, **0.010, *0.03 respectively). Specifically, except for #7 & #11 connection, the stressed group has weaker FC_P than the control group. This suggested that the callosal projection of the prefrontal area is weakened, whereas the callosal projection of the parietal area is strengthened by stress. However, the exceptional connectivity between #7 & #11 in this analysis is not further described, due to its low average FC_P value (control: -0.08, stress: 0.177), that indicates almost no connectivity exists between two areas. In the ipsilateral-hemisphere, Mann-Whitney U test of FC_P in the pair of #6 & #7 was diminished in stressed group (p<0.05; **0.003). The inter-cortical and ipsilateral connectivity is weakened and contralateral connectivity is strengthened in the stressed group compared to the control group.

In the recovery period (16~20s), three FC_P alterations were observed between control and stressed groups during 18~20s; #10 & #11, #9 & #14, #6 & #9 (Fig 3D). They were strengthened in the stressed group (*p<0.05, *0.0416, *0.020, *0.0416 respectively, Mann-Whitney U test). FC_P between seeds #10 & #11 and #9 & #14 are low in both control and stressed groups. Particularly, it was significantly diminished to the point where almost no connectivity was observed in some pairs of seeds (#10 & #11 control: 0.0094, stress: 0.2914), (#9 & #14 control: 0.0278, stress: 0.2787). Regarding a pair of seeds #6 & #9, a significantly increased FC_P at stressed group as moderately correlated to highly correlated relation. (control: 0.4295, moderately correlated / stress: 0.56236, highly correlated)

The graphical summary of the altered relationships (Fig. 3E~G) illustrates that under stress conditions, the posterior intra-connections of the contralateral hemisphere and inter-cortical connections are diminished during the stimulation period for Ca²⁺ responses (Fig. 3E). Additionally, in the stressed group, there is a reinforcement of contralateral and posterior inter-cortical FC_P and a weakening of inter- and ipsilateral cortical connectivity during the stimulation period for HbT (Fig. 3F). During the HbT recovery period, the stressed group shows weakened inter-cortical and ipsilateral connectivity, while contralateral connectivity is strengthened compared to the control group (Fig. 3G).

Spearman’s rank correlation was used to examine the relationship between ranks of Pearson correlation coefficients which was derived from a FC_P map between two cortical regions [Supp. 1]. Both the control group and the acutely stressed group displayed a strong correlation with each other, with correlation coefficients of r (36)=[.66126, .778157, .801928, .65537, .77851, .85089], and p-values of all the values ***p<0.001. Across all six relationships encompassing three different time points and two distinct signals, i.e., corrected Ca²⁺ and HbT responses, a robust positive correlation was found. Particularly, in the recovery period of the corrected Ca²⁺ fluorescence (ΔF/F) data, there was a strong correlation with rho > 0.8. The other five correlations also exhibited Spearman correlation coefficients exceeding 0.64. There was no significant rank change observed in the Pearson correlation coefficient values. In summary, when comparing linear correlation using Pearson correlation, differences in FC_P were observed. The stress mice group exhibited a pattern indicating impaired FC_P.

Thise study focused on investigating neural and hemodynamic changes in response to sensory stimulation in an acutely stressed group compared to a control group of mice. The mice were subjected to acute stress by immobilizing them, leading to higher corticosterone levels in the stressed group. The study analyzed calcium and hemodynamic signals in specific regions of the cortex under sensory stimulation. Our results revealed that acute stress impacts on brain-wide connectivity, affecting the FC between different cortical regions during stimulation and recovery periods. These FC alterations were evident in both calcium signals and hemodynamic responses, indicating complex changes in neural and hemodynamic signals under stress conditions.

The primary somatosensory (S1) barrel cortex exhibits reciprocal connections with other cortical areas, displaying primary connections with the secondary somatosensory cortex (S2) and the primary motor cortex (M1) within the same hemisphere. Our result showed that the FC within the previously known barrel projection pathway was strengthened during the stimulation period under acute stress. Overall, however, other routes, especially the frontal cortical inter-hemispheric connectivity, appeared to be weakened under acute stress. In addition, we observed a time delay in the recovery phase of responses in parts of the contralateral somatosensory, primary visual, and retrosplenial cortex. This delaying tendency is observed at both neural and hemodynamic response. These results suggest that the increased activity of glutamatergic transmission, induced by the release of glucocorticoid hormone in response to acute stress through the HPA axis, affected the excitatory pyramidal neuron that directs barrel projection. Our study is consistent with dynamic network analysis of acutely stressed human brains (Trier Social Stress Test) measured with fMRI showed the decrease of global-scale variability in the frontal-temporal regions and occipital pole prominently to be less-segregated state [24]. Our study utilizes ‘wide-field optical mapping’ method, which attains sufficient temporal and spatial resolution to capture neural activation and hemodynamic simultaneously in meso-scale as well as brain network connectivity in a wide-field of view.

Calcium and hemodynamic responses with a wide-field imaging under acute stress

Our calcium fluorescence and hemodynamic data is acquired simultaneously by a wide-field optical mapping [18]. Cerebral hemodynamic responses reflect neural activation levels via neurovascular coupling. The heightened neural activity, resulting in increased cerebral metabolic demands, triggers a corresponding increase in hemodynamic responses. In order to correct the possible crosstalk between calcium fluorescence signals and hemodynamic signals, we applied ‘single-wavelength method’ [18].

We acquired 10 images of reflectance and fluorescence per second respectively, meaning we obtained images at 100 milliseconds intervals. Even if temporal resolution of our implemented tool is not fully sufficient to capture all calcium signaling events in neurons, which occurs at millisecond time scale [25], a wide-field optical mapping is fast enough to detect widespread neural firing events as well as their dynamic coupling to hemodynamics [18]. It measures direct neural activity and temporal resolution is better than other perfusion imaging, such as fMRI. In addition, as the field of view is broader than single cell recording, we can analyze the overall brain activity not just capturing local activities from specific area.

Effect of Ketamine/Xylazine anesthesia to the stimuli response of cortex area

In the time lags with maximum cross-correlation at calcium fluorescence showed that recovery was delayed in acutely stressed group compared to the control group from seed #13. This may be related to Ketamine/Xylazine (K/X) anesthesia [17]. K/X anesthesia mouse affects the major structure of the visual pathway to evoke the bilateral significant blood-oxygen-level-dependent (BOLD) response compared to the awake mouse model. Under Ketamine anesthesia, widespread positive BOLD activity is disinhibited and excitatory neurons are sensitized. K/X anesthesia mouse showed delayed onset time, time-to-peak, FWHM compared to the awake group in V1 area. This effect of K/X anesthesia can lead to the visual area difference in stress condition, since its non-equivalent effect to each type of neuron [17]. Moreover, the whisker system likely has developed to compensate for the lack of visual information during the significant portion of a rodent’s life [23].

In the maximum cross-correlation data of hemodynamic reflectance showed that seed #13 cortex has slower recovery than other seeds and prominent post stimulation response in acutely stressed groups. But the control group also showed post stimulation response in seed #13. These post stimulation responses are reported to be observed in CBV-weighted (CBVw) optical intrinsic signal (OIS) responses and arterial dilation under the K/X anesthetized with spontaneous breathing conditioned mice model. This K/X anesthesia effects can be related to the retrosplenial cortex area difference in stress condition, since its non-equivalent effect to each type of neuron [26]. However, since we used the same K/X anesthetic in both the control and stressed mice groups, the likelihood of the effects we observed being due to the anesthetic rather than acute stress is low.

Limitations and future perspectives

We utilized a wide-field optical imaging to investigate excitatory neural calcium responses and hemodynamic responses under acute stress. To address the crosstalk between the calcium signal and hemodynamic signal in the calcium fluorescence data, we performed hemodynamic correction in the calcium fluorescence data. But, further study is needed to improve hemodynamic correction in the calcium fluorescence signal. Also, higher resolution imaging or molecular mechanism study are needed to answer for the differences observing in cortical regions with and without acute stress. Additional research for observing the recovery after the acute stress would be needed to confirm that disruption of brain network caused by acute stress can be restored. Of significance in our study is the network level assessment of the impact of acute stress on the whole cortical area.

This research was supported by the Institute for Basic Science (No. IBS-R015-D1); the National Research Foundation of Korea (NRF) funded by the Korea government (MSIT) (No. 2017R1A6A1A03015642, 2023R1A2C1004318, RS-2023-00302458); Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No. 2020-0-00261); a grant of the Korea Dementia Research Project through the Korea Dementia Research Center (KDRC), funded by the Ministry of Health & Welfare and Ministry of Science and ICT, Republic of Korea (No. RS-2024-00344426); the Fourth Stage of Brain Korea 21 Project in Department of Intelligent Precision Healthcare, Sungkyunkwan University (SKKU) (No. S-2022-0608-000); This research was also supported by KIST-SKKU Brain Research Center.