Chronic brain blood-flow imaging device for a behavioral experiment using mice

: A chronic brain blood-flow imaging device was developed for cerebrovascular disease treatment. This device comprises a small complementary metal-oxide semiconductor image sensor and a chronic fiber-optic plate window on a mouse head. A long-term cerebral blood-flow imaging technique was established in a freely moving mouse. Brain surface images were visible for one month using the chronic FOP window. This device obtained brain surface images and blood-flow velocity. The blood-flow changes were measured in behavioral experiments using this device. The chronic brain blood-flow imaging device may contribute to determining the cause of cerebrovascular disease and the development of cerebrovascular disease treatment.

with a small C flow and vess used to meas optical imagin no staining a blood was u hemoglobin [ flow systems Fig. 1 shows (c) sh pictur day af CMOS imaging sels are highly sure blood flow ng. Blood-flow and is minimal used. The illu 15,16]. The re in the brain rel 1. Chronic brain b s the outline of the hows a picture of re that captured in fter surgery. g device. Brain interested in b w for a long w imaging can lly invasive. T umination wav esult of this stu lated to drug p lood-flow imaging e device, and (b) s the mouse that ha n the day of surger n activities and brain studies [ time, because be used in clin The absorption velength was udy could cont otency for dise g device for a beh shows the chronic as the chronic FOP ry. (e) shows a pi d blood flow ar [12][13][14]. In this e the device re nical applicatio n spectrum of 535 nm, ef tribute to the u eases.
havioral experime c FOP window on P window on its icture that was cap re closely relate s study, this de equired no sta ons, because it the hemoglob ffectively abso understanding ent using mice. (a) the brain surface head. (d) shows a ptured in the 33th ed. Blood evice was aining for t requires bin in the orbed by of blood-

Chronic brain blood-flow imaging device for the cerebral blood flow of the mouse
The chronic brain blood-flow imaging device was developed for long-term mouse cerebral blood-flow measurement in a behavioral experiment. See Fig. 2(a). This device includes a CMOS image sensor, green light-emitting diode (LED) for light sources, and a fiber-optic plate (FOP) on a flexible printed circuit (FPC) (TAIYO INDUSTRIAL CO., Japan) in Fig.  2(b). The CMOS image sensor was designed and fabricated using the standard CMOS process (0.35-μm, 2-poly-4 metal CMOS; Austria Microsystems, Austria). The pixel size of the CMOS image sensor was 7.5 × 7.5 μm, and this sensor held 120 × 268 pixels. This device had six green LEDs (EPISTAR Corp., Taiwan) with an emission wavelength of 535 nm, located around the sensor as light sources. This wavelength is one of the absorption spectral peaks of hemoglobin in the blood. At this wavelength, blood flow is measured from the hemoglobin in the blood vessels. The LED size was 280 × 300 μm. Parts of the device were connected to other parts using a wire-bonding tool. Finally, an FOP (Hamamatsu Photonics, Japan) was mounted on the component side of the CMOS image sensor and LEDs. In this study, we used a high-resolution FOP J5734. The FOP is comprised of a bundle of micron-sized optical fibers. The diameter of the optical fiber was 3 μm. An incident image from an end face of the FOP was transmitted to the opposite side of the FOP. The FOP had the same optical quality as an optical fiber bundle for image transmission. In this study, alignment of FOP and CMOS was less affected, because the resolution of the CMOS image sensor (7.5 μm) was much larger than the resolution of the FOP (3 μm). In this study, we used a surface irradiation CMOS image sensor. When we mounted the sensor on the substrate, we wire-bonded the same surface of the CMOS image sensor. However, the CMOS image sensor did not closely attach to the chronic FOP window, because the imaging surface was not flat by the bonding wire. Therefore, we put the FOP on the pixel array area of the image sensor to make a flat surface. The FOP raises the imaging surface higher than the wire-bonding area. Because of this structure, the chronic brain blood-flow imaging device captured a clear brain surface image with a high spatial resolution and light intensity. Black paint was used to shield the side of the device (CANON CHEMICALS INC., Japan). The weight of the device was 0.40 g, 1/50 of an adult mouse. The power of the LED was 50-100 mW/mm. We adjusted light power for each experiment. The illumination uniformity was obtained by arranging 6 LEDs around the CMOS image sensor. The LED light was scattered onto the brain surface, because the brain tissue is a light-scattering material. The input and output signals of the image sensor were transmitted to a control board through a small relay board using four wires. The relay board contained an operational amplifier and a digital buffer used to regenerate the attenuated signals. The light intensity was controlled using a current generator through a small relay board. The control board included a digital-analog converter for transmitting signals to a personal computer (PC), because the sensor output contained analog signals. The analog output signals transmitted from the sensor were converted to 14-bit digital data. All the signals were controlled using an original PC program. In this experiment, the images were stored in the PC. Maximum frame rate was 120 Hz.

Chronic
A chronic FO chronic windo blood-flow im device does n out of focus u FOP was ma structure of th imaging devi was higher th brain-surface window was 6 core diameter 2.3 Animal p C57BL/6 mic animal proced Institute of Sc mg/kg of me introduced int fixed in a ste heater. The sk observed (pos performed ab sensory corte chronic FOP was placed on skull around between the s window on th determined th shape of the b FOP window 2. Structure of the ic brain blood-flow s the structure of , and a FOP on an FOP window OP window w ow uses thin g maging devices not have optica using this wind ade of glass o he FOP was us ce maintained han that of the images witho 6.78 x 6.86 x 2 r of the FOP fib preparation ce (SLC Co., J dures conform cience and Te edetomidine, 4 to the abdomin ereotaxic instru kull of the mou sterior 0.5 mm ove the observ ex in an area window was c n the brain surf the FOP after skull and ancho he head. For pu he position of blood vessels using the dent e chronic brain blo w imaging device the device. This n FPC.
for the chron was developed glass for the wi s cannot use th al components, dow. In this stu ptical fiber, a ed as the chron the spatial re CMOS image out direct cont 2.54 mm. The w ber was 3 μm. Japan, 7-10 w ed to the anim chnology. A c 4.0 mg/kg of nal cavity of th ument (Narishi use was expose and lateral 2.5 ved region. The corresponding constructed on face, and the an the craniotom ors. Figure 1(c utting the CMO the CMOS im on the brain s tal cement. used in this s n guidelines of was prepared w kg of butorpha 7]. The mouse se was kept w mary sensory co han 10-mm squ was positioned a imb regions. N n in Fig. 1

Blood-flow imaging and behavioral experiment
The blood-flow changes in the brain were obtained from changes in the reflection of the green light using the pixel value in the CMOS image sensor. The pixel-value data, stored in the PC, were analyzed by an original program written using Matlab. In the blood-flow velocity analysis, the velocity was shown as segments of the line scan along the blood vessel, plotted over time. The dark stripes corresponded to the red blood cells traveling along the length of the vessel, and the slope of the lines was proportional to the speed at which the hemoglobin traveled. We prepared the line-scan image with the distance, x (mm), of the measured blood vessel on the abscissa at time, t (s), of the scan of the ordinate. We calculated blood-flow velocity, V (mm/s), from V = x/Δt. Δt is the changing time of the slope of the line-scan image. We detected 10 lines from the line-scan image and calculated the average of the blood flow velocity (mm/s) from 10 lines. The range of detection of blood-flow velocity depends on the frame rate of the CMOS image sensor. Our image sensor operates within 1 to 120 fps. The range of the detection was 0.075 to 15 mm/s (measurement pixel number: 50; measurement frame number: 200). After obtaining the line-scan data, the data in the longitudinal area and the lateral lines were averaged to remove noise. For brain-activity measurement, the device measured brightness changes of the brain surface. Capillary blood vessels were not obtained, running throughout the brain surface. When brain activities occur, blood-flow volume of capillary blood vessels is increased temporarily for delivering sustenance and oxygen to neurons. In this experiment, we show the result of this. The value of Fig. 6(a) indicates the average of the pixel value rate (%) of the mean value, R, of the ROI (9 × 9 pixels). The pixel value rate was calculated from ΔR/R 0 . R 0 is the average value of the ROI in measurement time. ΔR was calculated from R-R 0 .

Conditions of the mouse cerebral surface with the chronic real-time imaging system
We observed conditions of the brain surface through the chronic FOP window after implantation. Figures 1(d) and (e) show pictures of the chronic FOP window captured on the day of surgery and 33 days afterwards. The brain surface kept good conditions one month afterwards. The chronic FOP window demonstrated use in long-time stable conditions. Figures 3(a) and (b) show the brain-surface images captured by a commercial camera. The brain surface had a small hemorrhage on the edge. However, the observation area had no damage, and the blood vessels were in good condition. The chronic FOP window had no effect on mouse behavior. The imaging device observed the same brain surface area in Fig.  3(c). The blood vessels appeared as black lines in this image, because the green light was effectively absorbed by the hemoglobin in the red blood cells.
In Fig. 4, the blood-flow velocity was measured with an imaging device. We set the device to a frame rate 66.41 fps and an exposure time 15.06 ms. The blood-flow velocity was analyzed by subtracting the images and the line scan images. In Visualization 1, the subtracted images made from the brain surface images. The blood flow was observed. An analysis of the blood-flow velocity was performed in a part of the line, as shown in Fig. 4(a). The results of the line scan are shown in Fig. 4(b). The direction of the slope was the direction of the blood flow, and the estimate of the slope shows the blood-flow velocity. The results of the blood-flow velocity are shown in Fig. 4(b). The blood-flow velocities of the vessel were estimated, as shown in Fig. 4(b). The velocities were 0.88 mm/s, 1.13 mm/s, and 1.22 mm/s.

Mouse device in a fr
For the freely head were im cortex. After imaging he mouse the visual ed on the brain surface. in Fig. 5(b). T enabled obser the blood flow we set the de brain-surface blood-flow m blood-flow ve Second, th M, 10-µL KC stimulation el brain activity were obtained brain surface measured the activity chang changes are sh In this experi The blue light rvation of bloo w was measure evice to frame image capture measurement is elocities in the he brain activit Cl solution was licited propaga at the brain su d by the imagin . The five pa light intensity ges were transm hown. 5. Cerebral blood e with the chronic iment, a cannula vioral experiment inated. A blue illu enetrate the skin ng experiment by t iment.
iment, the blue did not penet od flow in the b ed on the brain rate 110.64 fp ed in the freely shown in Fig  behavioral exp ty on the brain injected into t ation of Ca 2+ si urface. After inj ng device, as s rts surrounded y in Fig. 6(c). F mitted from po flow imaging in a c brain blood-flow was used to stim with the mouse umination light wa and tissues. (c) sh the device. (d) sho e light illumina trate into the s behavioral exp n surface with ps and exposur y moving expe g. 5(d). The im periments. The n surface was m the visual corte gnals in the br njection, the bra shown in Figs. d by square li Figure 6(c) sho oint 1 to point a freely moving ex w imaging device mulate the brain a . The green LED as used for this exp hows the brain su ows the line-scan i ated the experi skin and tissue periments using an imaging d re time 9.04 m eriment by the maging device e velocities wer measured by n ex using the ca rain, and this p ain activity on . 6(b) and 6(c) ines are the m ows light inten 5 Figure 5(c) device. The re obtained the m re 1.32 and 0.7 neural stimulat annula in Fig. 5 propagation cha the sensory co ). Figure 6( In the b e improved. T ame rete of 58 elocity from a brain blood-flo from a tiny blo ebral blood-flo riment, and the ent, the brain ated to brain a esults of the brain ) shows the brain ines indicate the m of the brain activ hows the graph of were transmitted fr ral blood flow revious study ding on the b to achieve lon chronic FOP w ethod similar t ent of the FOP observed throu minimally inva measurement o blood-flow im The imaging s 8 fps in the pr curved blood ow imaging de ood vessel with ow imaging exp e blood-flow ve n activities we activities in th n activity changes surface image cap measurement areas ity changes in the f the brain activity rom point 1 to poin w was difficult [10,11], bec brain surface. ng-term cerebr window techni to that of a typ P is glass. The ugh the FOP w asive. This a of blood flow maging method speed was upg revious study.
vessel. These evice. The CM h a diameter un periment, the b elocity was obt ere measured e sensory cort s in the brain sur ptured by the devi s of the changing e brain surface wit y changes at the b nt 5.
using the imp cause animal Therefore, a ral blood-flow ique produced pical chronic w FOP is not to window one mo good behavi for psychiatr ds, the imagin graded to 120 The analysis upgrades exp MOS imaging d nder 7.5 μm w blood flow was tained from mu using the de tex-were obta This device could simultaneously measure the blood-flow velocity and brain activity. These results suggested that the chronic brain blood-flow imaging device could be used for the development of a therapeutic methods for cerebrovascular disease [21]. The treatment effect and recovery process applying to cerebrovascular disease could also be determined by the devices ability to provide long-term observation in a freely moving experiment.

Conclusion
A chronic brain blood-flow imaging device was developed, and blood-flow measurement was demonstrated in a freely moving experiment using a mouse. This device comprised a small image sensor based on CMOS integrated circuit technology and a chronic FOP window on the mouse head. The chronic brain blood-flow imaging device obtained the blood-flow velocity of the cortex through the chronic FOP window. The chronic brain blood-flow imaging device observed clear brain surface images through the chronic FOP window 1 week after implantation. The blood-flow velocity and brain activity were obtained in the freely moving experiment using the device. The chronic brain blood-flow imaging device performed long-term brain imaging. The device can be used in long-term behavioral experiments for the development of therapeutic methods for treating cerebrovascular disease. The advantages of this device include its usefulness for experimentation, such as drug discovery research and disease research.