Behavioral tests in rodent models of stroke

Abstract

Rodents are the most widely used experimental animals in stroke research due to their similar vascular anatomy, high reproductive rates, and availability of transgenic models. However, the difficulties in assessing higher brain functions, such as cognition and memory, in rodents decrease the translational potential of these studies. In this review, we summarize commonly used motor/sensorimotor and cognition tests in rodent models of stroke. Specifically, we first briefly introduce the objective and procedure of each behavioral test. Next, we summarize the application of each test in both ischemic stroke and hemorrhagic stroke. Last, the advantages and disadvantages of these tests in assessing stroke outcome are discussed. This review summarizes commonly used behavioral tests in stroke studies and compares their applications in different stroke types.

1. Introduction

Stroke is the second most common cause of death and one leading cause of disability worldwide.¹ It is mainly categorized into ischemic stroke and hemorrhagic stroke, which are caused by occlusion and rupture of cerebral blood vessels, respectively. Hemorrhagic stroke is further categorized into intracerebral hemorrhage (ICH) and subarachnoid hemorrhage (SAH) depending on the location of hemorrhage. Among all stroke cases, ischemic stroke accounts for ~87%, while hemorrhagic stroke makes up ~10%.² Although less common than ischemic stroke, hemorrhagic stroke has a higher mortality rate (up to 67.9%) and worse prognosis.³ When stroke occurs, ischemia or hemorrhage in the brain impairs CNS function and causes a series of neurological defects, predominantly affecting the sensorimotor and coordination functions.⁴ In addition, studies have shown that stroke also leads to long-term cognitive impairment.^3-6

To enable stroke research and drug development, animal models of stroke have been established in several species, including rodents, pigs, and non-human primates.^4,7-10 Among them, rodents are most widely used due to high reproductive rates, low maintenance cost, availability of transgenic models, and similar vascular anatomy as humans.¹¹ Specifically, middle cerebral artery occlusion (MCAO) is commonly used to induce ischemic stroke,⁸ while intracerebral injection of collagenase or autologous blood is frequently used to induce ICH in rodents.^12,13 SAH can be induced in rodents simply by advancing suture through the external carotid artery to penetrate the vessel near intracranial bifurcation.^14,15 Accumulating evidence suggests that rodents exhibit multiple sensorimotor deficits and neurological impairments similar to human patients after stroke.^16-18 There is, however, one major challenge in rodent stroke models. Due to anatomical and functional differences between humans and rodents, it is difficult to evaluate higher brain functions, such as cognition and memory. Accurate assessment of neurological function and stroke outcome in rodents, which has translational applications, depends on the use of appropriate behavioral tests.

Here we review commonly used motor/sensorimotor and cognition tests in rodent models of stroke. Specifically, we first describe the objective and procedure of each behavioral test. Next, the application of each test in different stroke types and major findings are discussed. Last, the advantages and disadvantages of each test are summarized. This review aims to summarize commonly used behavioral tests in rodent stroke studies and compare their applications in different stroke models.

2. Motor and Sensorimotor Tests

2.1. Neurological Score

Neurological score was originally developed to evaluate stroke outcome in human patients. It assesses multiple aspects of neurological functions, including consciousness, vision, motor function, sensory function, verbal response, and brainstem reflexes.¹⁹ Various scoring systems, including the Bederson score, modified neurologic severity scores (mNSS), and Garcia scale, have been developed to evaluate stroke outcome in rodents.

2.1.1. Bederson Score

The Bederson score, which can be applied on both rats and mice,²⁰ evaluates global neurological function.⁸ The Bederson score grades rodents on a scale of 0-3. Grade 0 (no defects) is assigned if animals extend both forelimbs toward the floor when suspended above the floor. Grade 1 (mild defects) is given when animals show forelimb flexion without any other abnormality. Graded 2 (modest defects) is given when animals display forelimb flexion and decreased resistance to lateral push toward the paretic side. Animals further show circling behavior are scored Grade 3 (severe defects). To reflect more severe neurological deficits, the Bederson score system has been modified to incorporate Grade 4 and Grade 5. Specifically, animals with longitudinal spinning after stroke are scored Grade 4, and those without movement are scored Grade 5.²¹

The Bederson score has been widely used in ischemic studies. It is usually performed within 24 hours after surgery to evaluate acute stroke outcome. For example, rats showed circling behavior and were scored Grade 3 at 24 hours after MCAO.⁸ Similar result was observed in mice after ischemic stroke.⁶ Unlike in ischemic stroke, the Bederson score is less frequently used in ICH. This is because the Bederson score, as a simple assessment for locomotor function, is not sensitive enough for ICH. More sophisticated and sensitive tests are needed for outcome assessment in ICH. In addition, since the Bederson score mainly evaluates body asymmetry (e.g. circling behavior and unilateral limb deficits), it is not useful in SAH, which causes global brain injury.

The Bederson score does not require special equipment and is easy to perform. However, it has several limitations. First, it cannot accurately assess brain damage in certain regions. For example, rats with dorsolateral frontal cortex ablation failed to show circling behavior or defects in Y-maze performance.²² Next, the Bederson score is unable to evaluate long-term outcome because several deficits recover quickly after stroke. Third, the Bederson score cannot be used in models with global brain injury, such as global ischemia and SAH. Furthermore, the translational potential of the Bederson score is not high. This is mainly due to anatomical differences of human and rodent brains and the complexity of stroke outcome on humans.

2.1.2. mNSS

mNSS is another widely used neurological test in rodents after stroke. It combines neurological evaluations on multiple aspects, including motor function,^23-25 sensory function,²⁶ and reflex function,^27,28 with a total score of 14. One point is given when the animals fail to perform a task or show a reflex. A score of 1~4 indicates mild defects, a score of 5~9 suggests moderate defects, and a score of 10~14 indicates severe defects.

The mNSS has been used to assess long-term outcome in various stroke models. In a focal cerebral ischemia model, stroke rats demonstrated significantly higher mNSS score at days 1 and 28 after MCAO,²⁹ indicating neurological impairment. Interestingly, bone marrow stromal cell transplantation largely lowered the mNSS score at day 28 after stroke,²⁹ suggesting improved neurological function. Similarly, rats displayed a high mNSS score at day 1 after ICH,³⁰ indicating severe neurological impairment. Although recovered gradually, the neurological deficits remained detectable at days 24 and 42 after ICH.³⁰ In an SAH model, rats scored up to 16 points at day 1 after injury and 7-8 points at day 14 after injury.³¹ Mesenchymal stem cell transplantation significantly lowered the mNSS to 3 points at day 14 after injury compared to the controls,³¹ indicating improved recovery. These findings suggest that mNSS is a useful assessment for long-term stroke outcome and can be used to assess neuroprotective effects of novel treatments.

One advantage of mNSS is that it allows a comprehensive evaluation on multiple neurological functions, which makes it useful for long-term stroke outcome assessment. The multiple aspects, however, increase the complexity of mNSS test, making it less favorable. In addition, the combined score may cover some neurological deficits unique to certain stroke models. For example, some mice may get a high score on hindlimb motor function, while others may get a high score on forelimb motor function after hemorrhagic stroke. Due to body asymmetry and locomotor dysfunction, this difference may be covered by the similar overall score. Thus, individual score on each task should also be considered, especially in studies that investigate region-specific deficits.

2.1.3. Garcia Scale

To evaluate the functional outcome of MCAO, Garcia developed a neurological scale that assesses motor and sensory functions on rats.³² Specifically, animals are scored on their spontaneous activity (0-3), symmetry of limb movement (0-3), forelimb outstretching (0-3), climbing and grip strength (0-3), symmetry of body proprioception (0-3), and sensory function of vibrissae (0-3). The Garcia scale is represented as the sum of score in each test. In this scale, a lower score indicates more severe stroke outcome.

The Garcia scale is mainly used in ischemic stroke. In a transient MCAO model, rats got lower scores at day 7 after injury compared to sham controls.³² Similar result was found in a permanent MCAO model.³² In another study, mice displayed a low score at 22 hours after MCAO, which was elevated by Nrf2 activator.³³ These findings suggest that the Garcia scale is a sensitive assessment for motor and sensory functions after ischemic stroke. In addition, the Garcia scale has also been applied in other stroke models. In an ICH model, stroke mice exhibited a significantly lower score at 24 hours after injury compared to sham controls.³⁴ In an SAH model, stroke mice (11.2-12.4) scored much lower than sham controls (16.7-16.9) in the first 3 days after injury.³⁵ These results suggest that the Garcia scale can be used to assess neurological function in the acute phase after hemorrhagic stroke.

The Garcia scale is easy to perform, and assesses both sensory and motor functions. However, it should be noted that the Garcia scale focuses more on forelimb function than hindlimb function. In case that animals have severe hindlimb deficits but mild functional impairment on the forelimbs, the higher score in forelimb assessments will compromise the overall outcome. Thus, Garcia scale should be combined with other functional assessments to get a more comprehensive evaluation of stroke outcome. In addition, since the lesion site is controllable in ischemic stroke and ICH has a less predictable functional outcome, the Garcia scale is relatively more efficient in ischemic stroke.

2.2. Open field test

The open field test assesses locomotor ability and exploration behavior. It is widely used and has become a “standardize test” in rodent studies. The open field test was first developed on rats to study the open field behavior.³⁶ It was then modified and applied on mice.³⁷ Basically, a wooden or plastic open field maze with a size of 50cm (length) x 50cm (width) x 38cm (height) is used in this test. A mouse is placed in the open field and allowed to explore for 10 minutes. Its exploration behavior is recorded and analyzed through the video. Locomotor ability, the principle objective of this test, is revealed by the distance the mouse has traveled. A longer distance indicates a higher locomotor ability. By using automated analysis system, such as video tracking software, the route of the mouse can also be visualized and analyzed. Total travel distance and travel route are the most common parameters measured by the automated analysis system.³⁸ Additionally, the immobile time and rear up behavior can also be collected and analyzed to enable a comprehensive evaluation on locomotor ability.³⁸ The automated analysis system eliminates human errors and makes the result more objective.

In addition to locomotor ability, the open field test is also widely used to evaluate emotional behavior. On one hand, defecation and urination in the open field have been used as signs of emotionality.³⁶ However, defecation and urination are highly associated with the amount of food and water taken by the animals, which differs significantly among individuals. Additionally, strong emotions, such as anxious and scared, might decrease food intake, leading to reduced defecation. These make it relatively hard to standardize the signs of emotional behavior. On the other hand, the time animals spend in specific regions in the open field has been used to analyze anxiety level. Because rodents tend to stay at the corner or perimeters in a novel environment, the time they stay in the center of the open field reflects their anxiety level. The shorter time animals stay in the center of the open field, the higher anxiety level they have.³⁹

The open field test has been used to assess locomotor ability and anxiety level in ischemic stroke. Controversial results, however, have been observed. Many studies found no significant difference in travel distance in rats at days 3, 19, and 28 after MCAO.^40-42 In contrast, one report showed decreased travel distance at days 19, 28, and 42 after stroke.⁴³ This discrepancy may be caused by different rat lines. Although travel distance varies between studies, reduced rear-up behavior was commonly observed in rats from day 3 up to 6 weeks after MCAO,^40,41 indicating impaired locomotor ability. In addition, a hyperactive behavior was observed in rats 2-3 weeks after MCAO.⁴⁴ The hyperactivity may possibly be caused by hippocampal damage and spatial memory deficits, which slow down the habituation process and prolong the novel stimulation.⁴⁵ Furthermore, the open field test has also been used to determine anxiety level after ischemic stroke. It was reported that mice spent significantly less time in the center 48 hours after MCAO.⁴⁶ Similarly, rats spent longer time in the corner after transient forebrain global ischemia,⁴⁷ indicating an acute increase of anxiety level. It should be note that multiple factors, including habituation to environment and size of center area, may affect the anxiety-like behavior. Therefore, other anxiety tests, such as elevated-plus maze, should be used to more objectively evaluate anxiety level.

The open field test has also been used in hemorrhagic stroke. In an ICH model, rats showed significantly reduced travel distance at day 1 after injury compared to sham controls,⁴⁸ indicating decreased motor function. The motor impairment gradually recovered and was undetectable by day 14 after ICH.⁴⁸ Consistent with this finding, ICH rats showed no difference in open field test 2 and 8 weeks after injury compared to sham controls.^48,49 Unlike locomotor ability, the anxiety-like behavior after ICH is rarely studied by open field test. In an SAH model, rats exhibited declined travel distance at day 21 after injury and spent less time in the center,⁵⁰ indicating chronic motor impairment and increased anxiety. Similar results were observed on SAH mice at days 13 and 27 after injury.⁵¹ These results suggest that the open field test can be used to evaluate locomotor ability and anxiety after hemorrhagic stroke.

The open field test is an easy test and does not require sophisticated equipment. However, it should be noted that the open field test has various protocols and many factors may affect the result. For example, while 10 minutes is commonly used as the exploration duration, some might apply shorter time, such as 2 minutes. In this case, the test will focus more on the exploration behavior that responds to novel environment instead of locomotor ability. Moreover, when testing disease models that cause locomotion deficits, a dark environment might be needed to promote the exploration, because light may further decrease the locomotion activity of rodents. Recording in a dark environment, however, requires a higher standard of videoing device, which increases the difficulty of the test.

2.3. Pole Test

The pole test is an easy test to assess the overall locomotor function in rodents. It was initially developed on mice to study bradykinesia and later adapted to stroke studies.^52,53 Although the pole test has been used in rats for other studies,^54,55 it is seldom used in rat stroke models. This is possibly because rats are heavier than mice, which makes them unable to perform the task after stroke. In this test, a mouse is placed on the tip of an 8mm (diameter) x 50cm (length) wooden pole with its head upward (Figure 1A). The mouse then tries to descend to the floor by turning its head downward (Figure 1A). The latency to make the turn (T_turn) and the time to descend (T_D) are recorded. For mice that descend without turning their heads downward, the T_D is used to represent T_turn. If mice make the turn but fall in the halfway when descending, the total time until they reach the floor is recorded. If mice fall immediately, the maximum duration of 120s is assigned to the T_turn and T_D. It should be noted that mice cannot pause during descending. If this happens, the trial should be excluded and repeated. For better performance, mice need to be trained before stroke induction.

Fig. 1.

Illustration of commonly used behavioral tests in rodent models of stroke. A. Pole test. B. Foot-fault test. C. Cylinder test. D. Corner test. E. Rotarod test. F. Single pellet reaching task. G. Morris water maze. H. Y maze. I. Novel object recognition/location tests. J. Radial-arm maze. K. Elevated-plus maze.

The pole test is a useful assessment of locomotor function in ischemic stroke. It was reported that ischemic mice showed increased T_turn and T_D compared to sham controls up to 11 days after injury,^4,56 suggesting locomotor impairment. By day 17 after stroke, however, similar T_turn and T_D were observed in sham and ischemic mice, indicating recovery.^4,56 In addition, fluoxetine was able to improve animal performance in the pole test at days 12 and 20 after stroke,⁵⁷ suggesting that the pole test can be used to assess neuroprotective effects of novel treatments.

In addition, the pole test has also been applied in hemorrhagic stroke. In an autologous blood-induced ICH model, mice spent more time descending compared to sham controls at day 2 after stroke, and dexmedetomidine treatment significantly improved their performance.⁵⁸ In contrast, another study showed that mice failed to exhibit locomotor deficits in the pole test at days 2-7 after collagenase-induced ICH.⁵⁹ These controversial results might be caused by different ICH protocols. In SAH models, however, the pole test is seldom performed.

The pole test is relatively easy to perform and requires minimal equipment. It is an objective and sensitive test for motor function. Furthermore, the pole test is also able to assess long-term motor function in ischemic stroke.

2.4. Foot-Fault Test

The foot-fault test, also known as grid walking test, assesses motor function and limb coordination in rodents.^60-62 Basically, animals are placed on an elevated grid with square openings (1.69 cm² for mice and 6.25 cm² for rats) and allowed to move across the grid (Figure 1B). They move around the grid by placing feet on the wire frame. If the paw falls from or slips off the frame, one foot-fault is recorded. The total number of steps it takes to cross the grid and the foot fault for each limb are quantified. Sham controls have few foot faults with no bias toward either side,⁶³ while mice with stroke display increased foot-faults toward the contralateral side.⁶² To reduce individual variation, a pre-stroke test should be performed to customize the post-stroke result.

The foot-fault test has been used to study limb coordination in ischemic stroke. Short-term studies revealed that mice made significantly more foot-faults at day 2 after MCAO compared to sham controls,^62,64 indicating acute coordination deficits. The coordination deficits can be observed up to 90 days after stroke.⁶² Compared to wildtype controls, mice with attenuated astrocyte reactivity displayed substantially more foot-faults up to 4 weeks after MCAO, indicating slower recovery.⁶⁴ Similarly, the foot-fault test is also able to detect short-term and long-term deficits in ischemic rats. It was reported that increased foot-faults were observed in rats at days 2-28 after MCAO.⁶⁵ In addition, rats exhibited increased foot-faults in a hypoxia–ischemia model and the motor coordination deficits could still be detected 5 weeks after surgery.⁶⁶ These findings suggest that the foot-fault test is sensitive enough to detect both short-term and long-term motor coordination deficits in ischemic stroke.

The foot-fault test has also been used to detect limbs deficits in hemorrhagic stroke. Increased foot-faults were observed in mice at days 3 and 7 after ICH,^67,68 indicating short-term coordination impairment. Similar results were reported in ICH mice 2 and 3 weeks after stroke,⁶⁹ indicating long-term coordination deficits. Like mice, rats also displayed limb deficits at both acute (day 1) and chronic (day 21) phases after ICH.⁷⁰ In addition, rats also showed increased foot-faults 2 weeks after SAH.⁷¹ These results suggest that the foot-fault test is also useful in the assessment of motor coordination deficits in hemorrhagic stroke.

The foot-fault test is effective and objective in assessing motor function and limb coordination. Similarly, it can also evaluate long-term stroke outcome in both ischemic and hemorrhagic models.

2.5. Cylinder Test

One of the most significant symptoms of stroke is sensorimotor asymmetry, which can be assessed by many behavior tests, including cylinder test. Cylinder test was initially used to evaluate spontaneous motor activity of the forelimbs after CNS injury.⁷² It was modified to test limb-use asymmetry after stroke.⁷³ Generally, animals are placed in a clear cylinder container (9cm diameter x 15cm height for mice, and 20cm diameter x 30cm height for rats) and their exploring behavior (use of forelimbs to contact the wall and land) is recorded with a camera (Figure 1C). A mirror is placed behind the cylinder to enable recording when animals turn away from the camera. A total of 20 limb movements or a maximum of 10 minutes are recorded. By quantifying the percentage of impaired and non-impaired forelimb use for exploring and landing, limb-use asymmetry can be assessed. Since rodents are less likely to move after certain injuries, including stroke, this test can be performed in dark environment to encourage locomotion.

The cylinder test has been used to determine limb use asymmetry in ischemic stroke. It was reported that mice used ipsilateral forelimb much more frequently than contralateral limb after MCAO.⁷⁴ The asymmetry peaked at day 3 after injury, remained detectable at day 15 after injury, but disappeared by day 40 after injury.⁷⁴ In addition, mice with hypoxic ischemia showed more ipsilateral limb movements compared to sham controls at days 10 and 21 after injury,⁷⁵ indicating long-term limb asymmetry. Similarly, rats with focal ischemia displayed higher ipsilateral forelimb usage than pre-surgery baseline level up to day 28 after injury.^76,77 Interestingly, bone marrow cell treatment significantly enhanced the recovery of ischemic rats, which exhibited no limb use asymmetry in cylinder test at day 28 after injury.^76,77 Furthermore, in a long-term cortical ischemic study, ischemic rats demonstrated a strong asymmetry even at day 77 after injury.⁷⁸ These results suggest that the cylinder test is a sensitive assessment for long-term limb asymmetry in ischemic stroke and can be used to evaluate neuroprotective effects of novel treatments.

In addition, the cylinder test can also be used in hemorrhagic stroke. In the collagenase-induced ICH model, rats showed increased ipsilateral limb usage up to 4 weeks after stroke.^79,80 Interestingly, rehabilitation treatment was able to improve the performance of ICH rats in this test at day 28 after stroke, indicating faster recovery.⁸⁰ Similar results are also observed in autologous blood-induced ICH model. For example, rats showed elevated ipsilateral forelimb usage at day 28 after injury, and oxygenase inhibitor significantly improved the outcome by restoring contralateral forelimb usage to baseline level at day 14 after stroke.⁸¹ In sharp contrast to ICH, SAH rarely induces limb asymmetry in the cylinder test,⁸² probably because SAH causes global injury rather than unilateral damage in the brain.

The cylinder test is an easy and objective test for behavioral asymmetry. No special equipment and little administrative requirements are needed for this test. The cylinder test is able to detect long-term stroke outcome in both ischemic stroke and ICH models.^76-80 However, the cylinder test is not useful for models with global brain injury, such as SAH.

2.6. Corner Test

The corner test assesses the direction pattern of sensorimotor dysfunction. It was initially applied on rats,⁸³ and later modified to use on mice.⁶² Generally, two 30cm × 20cm cardboards are attached at 30° angle with a small opening at the joint (Figure 1D). Animals are placed in the middle of the two cardboards, so that vibrissae on both sides can be stimulated when they enter the corner (Figure 1D). The direction in which they turn back is recorded. Normally, animals do not have preference to either direction. After stroke, which causes contralateral limb deficits, however, animals tend to use their ipsilateral limbs to turn back. Higher tendency to one side indicates more severe stroke outcome.

As a sensorimotor asymmetry assessment, corner test has been applied in unilateral stroke studies to evaluate stroke outcome. Ischemic mice made more ipsilateral turns, while sham controls turned to either side equally after ischemic injury.⁶² This behavioral asymmetry was detected from 2-90 days after injury.⁶² Similarly, ischemic rats displayed significantly more ipsilateral turns than sham controls after ischemic stroke, and the difference could still be observed 4 weeks after injury.⁸⁴ In addition, ischemic rats with bone marrow cell treatment made fewer ipsilateral turns within 28 days after injury,^76,77 indicating decreased sensorimotor deficits. These results suggest that corner test can be used to assess sensorimotor asymmetry after ischemic stroke and evaluate neuroprotective effects of novel treatments.

The corner test is also effective in assess sensorimotor function in hemorrhagic stroke. It was reported that ICH rats showed significant increased ipsilateral turns compared to sham controls, and that this asymmetry remained detectable 4 weeks after injury.⁸⁵ In addition, ICH rats with deferoxamine treatment made less ipsilateral turns at day 28 after stroke, indicating improved recovery.⁸⁶ These findings suggest that the corner test is also able to assess sensorimotor function in ICH. It should be noted that the corner test cannot be used in SAH, which causes global brain damage and thus minimal sensorimotor asymmetry.

Compared to other tests that assess sensorimotor asymmetry, corner test has several advantages. First, it is an objective and quantitative test. Next, the corner test is able to evaluate sensorimotor deficits in the chronic phase after stroke.⁸⁷ One drawback of the corner test, however, is that animals may not perform well when they are too sick or lose motivation due to repeated testing.⁸⁸ To reduce or avoid this effect, long intervals should be given between tests. In addition, the use of corner test should be limited to unilateral stroke models, since global stroke models (e.g. global ischemia and SAH) result in mild sensorimotor asymmetry.

2.7. Adhesive Removal Test

The adhesive removal test is widely used in rodents to evaluate sensorimotor dysfunction and motor asymmetry.^4,73,89-91 Briefly, animals are placed into a 15cm x 25cm transparent box and two similar adhesive tapes are attached to the hairless part of each forepaw with the same pressure. The time it takes to contact and remove the stimuli is recorded. In general, animals spend more time contacting and removing the adhesive tape from the contralateral forepaw, while they have no problem contacting and removing the adhesive tape from the ipsilateral forepaw.⁴ In addition, adhesive removal test with modified protocols can also be used in stroke studies. For example, adhesive tapes were attached on the vibrissae of rodents to study sensory function in a photothrombotic stroke model,⁹² and adhesive tapes of various sizes were used to determine the stimulating threshold for removal behavior in a focal ischemic study.⁷³

The adhesive removal test has been used to assess long-term stroke outcome in ischemic models. In a distal focal cerebral ischemia model, the adhesive removal test was able to detect sensorimotor deficits in mice up to 3 weeks after injury.⁹³ Another study showed that cortical ischemic rats spent significantly more time removing the stimuli on forelimb wrists compared to sham controls 6 weeks after injury.⁹⁴ Using a modified protocol, it was reported that larger sticky tapes were required to trigger the removal behavior on contralateral forelimb (function impaired side) in ischemic rats in both acute (1-14 days after injury) and chronic (21-30 days after injury) stages in a focal ischemia study.⁷³ In addition, cerebral ischemic rats with bone marrow cell treatment exhibited reduced time to remove the stimuli at day 14 after injury,⁹⁵ indicating faster recovery. Together, these results suggest that the adhesive removal test is a sensitive test for long-term function assessment after ischemic stroke and can be used to evaluate neuroprotective effects of novel treatments.

The adhesive removal test has also been used to assess sensorimotor function in hemorrhagic stroke. Mice with ICH showed a longer latency to contact the stimuli and took more time to remove it,⁹⁶ indicating sensorimotor impairment after stroke. This difference could be detected up to 21 days after injury and bone marrow cell treatment significantly improved the sensorimotor function.⁹⁶ Similarly, rats exhibited a longer latency to contact the stimuli and/or spent more time removing the sticker at day 21 after SAH.⁹⁷ Additionally, SAH rats perforated from the right internal cerebral artery (ICA) showed more significant sensorimotor deficits on left paw than right paw at day 6 after stroke, and these deficits were attenuated by intranasal stem cell treatment.⁹⁸ These findings suggest that the adhesive removal test can be used to assess long-term outcome and recovery pattern in hemorrhagic stroke.

The adhesive removal test has various advantages. It is an objective and sensitive test for sensorimotor function assessment, and it can be used at the chronic phase after stroke to assess long-term stroke outcome. However, it has large variations and requires multiple rounds of training and strict control of the variables to obtain reliable data. For example, animals are usually trained once a day for 5 days before stroke induction to minimize individual variation.

2.8. Rotarod Test

The rotarod test evaluates equilibrium behavior and locomotor ability in rodents.^99-101 It has been used in many disease models, including stroke.¹⁰² Basically, a 3cm (diameter) x 40cm (length) rod is covered with sticking plasters to increase roughness. An electric motor is applied to rotate the rod at the speed of 20rpm. A landing platform with soft surface is placed 18cm below the rod to protect falling animals. Before the test, animals are trained to perform the task. They are placed on the rod with no rotation for 30 seconds, then with a constant slow (4rpm) rotation. Animals should be trained until they can stay on the rod for at least 1 minute. After the training trial, animals are placed on the rod for the testing trial, in which the rod rotates with an increasing speed from 4rpm to 20rpm over 2 minutes (Figure 1E). The latency before they fall on the platform is recorded. Rodents with impaired locomotor function fall from the rod faster than those with normal locomotor function.¹⁰¹

The rotarod test has been used to assess locomotor function after ischemic stroke. While rats were able to stay on the rod throughout the whole task after training, those with cerebral ischemia fell out quickly compared to sham controls, and the difference remained significant even after 6 weeks.^94,95 Interestingly, ischemic rats receiving treatments, such as neurotrophic factors and bone marrow cells, showed improved long-term recovery up to 35 days after injury.^94,95 These findings suggest that the rotarod test is capable of detecting long-term outcome and recovery pattern in ischemic stroke.

In addition, the rotarod test has also been applied in hemorrhagic stroke. In an ICH model, stroke mice had an acute decrease of latency in the rotarod test at day 3 after injury compared to sham controls.⁹⁶ A gradual recovery pattern was observed at days 3-21 after injury, and it was improved by stem cell treatment.⁹⁶ Consistent with this report, rats with neural stem cell transplantation exhibited improved performance in rotarod test at day 60 after ICH.¹⁰³ Similarly, SAH mice showed reduced latency before falling compared to sham controls at 24 hours after injury.¹⁰⁴ In another study, SAH rats displayed faster falling in the rotarod test at 48 hours after injury, and Nrf2 activator improved their performance.¹⁰⁵ A long-term study found motor deficits in SAH mice 3 weeks after injury.¹⁰⁶ These studies suggest that the rotarod test is able to detect long-term outcome and recovery pattern in ICH and SAH.

The rotarod test is a sensitive, objective, and quantifiable test. It is capable of detecting equilibrium behavior and locomotor function at both acute and chronic phases after stroke. It can also be used to determine recovery pattern and evaluate neuroprotective effects of novel treatments. However, the rotarod test requires special equipment and training is usually needed to obtain consistent results.

2.9. Wire Hanging Test

Wire hanging test assesses multiple aspects of locomotor ability, including grip strength, endurance, and body corporation. It is widely used in rodents with neurological disorders and/or muscle weakness.^107-109 Basically, animals are placed on a wire hanging 50~60cm above the ground for a maximum of 4 minutes, so that they have to suspend their bodies with limbs. The time animals spend on the wire (latency before falling), which reflects muscle strength, is recorded. While hanging, animals might use forelimbs or all four limbs to hold the wire. The different ways of hanging affect their performance, creating variation. To obtain more objective and reliable results, strategies that limit the ways of hanging, such as covering hindlimbs with adhesive tape, may be applied.¹¹⁰ In addition to muscle strength, the wire hanging test can also evaluate body corporation ability. In this case, different hanging behaviors are scored as described below: 0, animals directly fall off the wire; 1, animals hang with forelimbs; 2, animals hang with forelimbs and attempt to climb up; 3, hindlimbs are involved in hanging; 4, animals hang with four limbs and tail; and 5, animals successfully escape.¹¹¹ This scoring system correlates well with the locomotor ability and body corporation ability. Moreover, a modified hanging test that uses a vertically hanged wire hoop has been developed to more objectively assess muscle performance and endurance.¹¹² The hoop rolls as animals climb, keeping them in the same position. Animals tend to move consistently in one direction with similar behaviors.

The wire hanging test has been used to assess locomotor function in rodents after ischemic stroke. It was reported that rats failed to stay on the wire until day 5 after hypoxic ischemia, and the latency to fall gradually increased overtime and became fully recovered on day 17 after stroke.^113,114 Additionally, minocycline was able to improve their performance in the wire hanging test during the recovery phase.^113,114 Similarly, an acute decrease of hanging latency was observed on mice after ischemic stroke.¹¹⁵ However, unlike rats that fell off immediately during the first 5 days post stroke, mice were able to hang on the wire 24 hours after stroke.^114,115 This difference might be caused by their different body weights. In a short-term study, the wire hanging test revealed locomotor differences between mice receiving rosuvastatin treatment and controls at days 2, 3, and 5 after cerebral ischemia.^115-117 A long-term study also demonstrated that ischemic mice had less latency on wire up to 3 weeks after injury.¹¹⁸ These findings suggest that the wire hanging test is able to evaluate long-term locomotor function after ischemic stroke and assess neuroprotective effects of novel treatments.

In addition, the wire hanging test has also been used to evaluate locomotor outcome in hemorrhagic stroke. In a cortical ICH model, mice showed decreased falling latency after injury, which recovered to the baseline 2 weeks after stroke.¹¹⁰ It should be noted that mice with hippocampal ICH and sham controls exhibited comparable falling latency,¹¹⁰ suggesting that hippocampal impairment hardly affects the hanging ability and grip strength. Interestingly, fingolimod, a sphingosine-1-phosphate receptor modulator, substantially increased the latency on wire at days 1 and 3 after ICH.¹¹⁹ Similarly, in an SAH model, mice displayed locomotor deficits at days 1, 3, and 14 after surgery, indicating both short-term and long-term locomotor deficits.^71,120 In addition, cyclooxygenase-2 inhibitor increased the falling latency of SAH mice, indicating improved locomotor function.¹²⁰ These results suggest that the wire hanging test is useful in the assessment of long-term hemorrhagic stroke outcome and potential neuroprotective effects of novel treatments.

The wire hanging test is relatively easy to perform and has low requirements on equipment. It can assess both short-term and long-term stroke outcomes. However, training is required to reduce individual variation and obtain consistent results.

2.10. Skilled Reaching Tasks

Based on the association of limb function and cortex region,¹²¹ skilled reaching tasks were developed to assess limb motor function. They are widely used to evaluate functional outcome after stroke.¹²² The skilled reaching tasks are a series of tasks that train animals to reach their limbs through limited space, such as slots, staircases, or tubes.^123-125 Food pellets are commonly used to trigger the reaching behavior. Single pellet reaching task is a simple but widely used form of skilled reaching tasks. It was initially developed on rats and later adapted to mice.^123,126 In this test, an animal is placed in a plexiglass reaching box (19.5cm x 8cm x 20cm for mice and 30cm x 14cm x 45cm for rats) with a 10mm-wide slot in the middle of the front wall (Figure 1F). A tray is placed in front of the slot with two indentations spaced 1cm away from the slot on both sides (Figure 1F). The animal first receives a training trial to learn how to obtain food pellets. In the training trial, a food pellet is placed in the tongue distance on the tray, and then gradually moved farther so that the animal needs to use its paw to retrieve the pellet. An effective reaching can only be performed when the pellet is placed in the contralateral side of the preferred limb.¹²³ After a week of training, animals learn how to retrieve food pellets. Then, pellets are placed in both indentations to reveal animal’s preferred limb. Twenty pellets are given each day. A successful reach is counted when the animal obtains the pellet and a failure is counted if the animal knocks the food away or drops the pellet. Unilateral brain damage can be assessed by the percentage of successful reaching to the pellets on the contralateral side of the reaching limb.

The single pellet reaching task can be used to distinguish compensation and recovery after unilateral neurological injury, such as stroke.¹²⁷ Compensation can develop after stroke to make up the functional impairment, which may interfere with the evaluation of recovery patterns.¹²⁸ The single pellet reaching task is able to assess motor functions of ipsilateral and contralateral limbs independently. Enhanced performance of unimpaired limb may be a sign of compensation instead of recovery.

The single pellet task has been used to assess skilled motor recovery in various ischemic models. Rats with ischemic insult on caudal forelimb area showed a significantly lower success rate of reaching task, which gradually recovered in 10 days after injury.¹²⁹ Similarly, rats with ischemic stroke in the sensorimotor cortex exhibited poor performance in the single pellet task.¹³⁰ Compared to those without rehabilitation, which had a low recovery rate and severe skilled motor deficits, ischemic rats with rehabilitation training recovered faster and showed pre-stroke-level performance in the single pellet task 5 weeks after injury.¹³⁰ In addition, focal ischemic stroke also dramatically reduced the success rate in the single pellet task.¹³¹ It was reported that skilled motor deficits remained at day 28 after stroke and mice treated with bone marrow stromal cells displayed increased recovery 1 week after injury.¹³¹ These results suggest that the single pellet reaching task is able to evaluate long-term skilled motor deficits in ischemic stroke and assess potential neuroprotective effects of novel treatments.

The single pellet task has also been applied in hemorrhagic studies. In an ICH model, rats showed skilled motor deficits at both days 7-11 and days 28-32 after stroke with relatively better performance in the latter time point.¹³² ICH mice also demonstrated significantly decreased reaching success up to 28 days after stroke.¹³³ These findings suggest that the single pellet reaching task is also suitable for skilled motor deficit assessment in hemorrhagic stroke. It should be noted, however, that the severity of ICH may affect the outcome of this test. A study reported that rats with severe ICH failed to perform the task.⁷⁹

Skilled reaching tasks, such as the single pellet task, are specified for assessing skillful use of forelimbs and detecting skilled motor deficits. The single pellet task is a sensitive test, and can be used to evaluate both short-term and long-term stroke outcomes. However, a series of training trials are required to perform these tasks. In addition, individual variants largely affect the performance of animals.

3. Cognition tests

Cognition involves a variety of important neurological functions, including memory and emotions. Many behavioral tests have been developed to assess memory in animals. In primates, such as humans and monkeys, visual recognition is commonly used to reflect memory.^134,135 Specifically, the subjects need to identify familiar and unfamiliar visual stimuli compared to the stimuli that have been displayed before. This is, however, too complex for rodents due to the physical differences between species. Thus, tasks that follow their natural responses have to be used in rodents. A series of behavioral tests, including Morris water maze, elevated-plus maze, Y-maze, novel object recognition/location tests, and radial-arm maze, have been developed to assess memory and cognition in rodents.

3.1. Morris Water Maze

Morris water maze test assesses both cognition and locomotor function. It is widely used to evaluate long-term cognitive function after stroke in rodents.^136,137 In this test, animals are placed in a large (120cm diameter for mice and 180cm diameter for rats) circular water pool and forced to swim (Figure 1G). The pool is divided into four quadrants, and a platform is hidden just below the water in one quadrant (Figure 1G). The water is colored with paint to prevent visualization of the hidden platform. Animals can escape once they find the hidden platform. The latency to reach the platform is used as a readout to reflect the cognitive function. Increased latency indicates memory deficits and cognitive decline. In addition to latency to reach platform, the time animals spend in the target quadrant may also be used as a readout.¹³⁸ With the help of automated video tracking software, the route of animals, total swimming distance, and average swimming speed can also be visualized and analyzed.

Morris water maze has been used to assess cognitive function after ischemic stroke. It was reported that mice with focal ischemic stroke showed increased latency to find the hidden platform 2, 4, and 6 weeks after injury compared to sham controls.^139,140 Similar result was observed in rats 12-14 weeks after ischemic stroke,⁹⁰ indicating declined cognitive function in rodents after ischemic stroke. In addition, rats that stayed in an enriched environment spent less time finding the hidden platform than those in deprived environment 5 weeks after ischemic stroke,¹⁴¹ suggesting a beneficial role of enriched environment in cognitive function after stroke. It should be noted, however, that there is also evidence showing unaltered cognitive function in mice after stroke. For example, mice with and without stroke failed to exhibit any difference in the latency to find the hidden platform 2 weeks after MCAO.⁴ This discrepancy may be due to different protocols used in these studies.

In addition, Morris water maze has also been used to evaluate cognitive outcome in hemorrhagic stroke. In an ICH model, rats displayed increased swimming distance 2 but not 8 weeks after injury compared to sham controls.⁴⁹ In this study, the swimming distance instead of latency was used as a readout for spatial memory. This is because the swimming speed may interfere with the interpretation of the result. For instance, animals with higher swimming speed may find the platform faster even with a poorer spatial memory. Using swimming distance eliminates the effect of different swimming speeds and more accurately reflects the spatial memory of tested subjects. Consistent with this finding, another study also reported that ICH rats failed to show any difference in the Morris water maze 8 or 16 weeks after injury.¹⁴² These results suggest that Morris water maze is able to detect cognitive impairment in the subacute phase (~2 weeks) after ICH. In an SAH model, SAH rats displayed significantly increased escape latency at day 5 but not days 1-4 after stroke compared to sham controls.¹⁴³ In another study, SAH rats displayed spatial learning deficits at days 4 and 5 after SAH,¹⁴⁴ indicating short-term cognitive impairment. In a long-term study, SAH rats exhibited prolonged escape latency at days 29-35 after stroke,¹⁴⁵ indicating long-term cognitive decline. The Morris water maze is less frequently used in mice than in rats in SAH. A recent study revealed that SAH mice spent longer time to locate the hidden platform during learning trials on days 1-4, and less time in target quadrant during probe trial on day 5 compared to sham controls.¹⁴⁶ Similar result was observed in SAH mice on days 14-19 after injury.¹⁴⁷ These results suggest that Morris water maze is able to detect SAH-induced spatial learning deficits in rats up to 5 weeks after injury and in mice up to 19 days after injury.^143-147 Future research should determine if Morris water maze is able to assess cognitive function in SAH mice at later time points.

Morris water maze allows assessment of both cognition and locomotor function in rodents. It is useful in the evaluation of both short-term and long-term outcomes of stroke. This test, however, requires long training trials, and the training protocols may affect animal performance. Additionally, its results are affected by the locomotor ability of animals. Therefore, it is better to use this test for animals with similar swimming speed. Alternatively, modifications may be applied to obtain more reliable results. For example, swimming distance before reaching the platform rather than time spent in target quadrant may be used in case that animals have different locomotor ability.

3.2. Y-maze

The Y-maze is commonly used to assess spontaneous alternation and spatial memory in rodents.¹⁴⁸ It contains three arms with 120° angle between each other. The Y-maze was initially used to study exploration behaviors^149,150 and behavior pattern (known as spontaneous alternation).¹⁴⁹ Later, the Y-maze was used to assess spatial memory in rodents by evaluating spontaneous alternation.¹⁴⁸ This test includes a single trial, in which the animal is placed in one arm of the Y-maze and allowed to explore for a period of time (Figure 1H). The exploration time varies from 3 to 15 minutes depending on the types of research.^148,151-153 The sequence of each entrance and total number of entrances are recorded. Spontaneous alternation is shown by the tendency of the arm visits that are different from previous two visits. For example, if the three arms are represented as arms 1-3 and the sequence of animal visits is (1,2,3,2,1,3,1,3), the total alternation opportunities would be all the visits except the first two, which is six. The third, fifth, sixth visit were different from previous two, so the percentage of alternation is three in six (50%).

Because rodents get habituated to the environment, long-time exposure to the maze environment may decrease their exploration activity and affect their alternation tendency.¹⁴⁹ To minimize those factors, a two-trial memory task using Y-maze was developed based on the single trial Y-maze.¹⁵⁴ In the first trial, one of the arms is blocked and the animal is allowed to explore the rest two arms (Figure 1H). In the second trial, the blocked arm is opened and considered as the novel arm. The animal is allowed to explore all three arms and the entrance to each arm is recorded. The spatial memory of animals can be revealed by a discrimination to the novel arm in the second trial. It should be noted that the exploration time and inter-trial interval duration may also affect the result. It was reported that animals exhibited higher discrimination to the novel arm in the first 3 minutes of exploration with 30 minutes interval.¹⁵⁴

The Y-maze task has been used as a cognitive assessment after ischemic stroke. In a focal cerebral ischemic model, the spontaneous alternation of ischemic rats (67.1%) was slightly less than sham controls (76.0%) at day 3 after injury, indicating decreased short-term cognitive function.⁴¹ Similarly, reduced spontaneous alternation in Y-maze was also observed in mice at days 7-13 after ischemic injury.¹⁵⁵ In a long-term study, the same result was found in rats 8 weeks after focal cerebral ischemia.¹⁵⁶ These findings suggest that the Y-maze can be used to assess cognitive function at both short-term and long-term after ischemic stroke.

Unlike in ischemic stroke, the Y-maze test is less frequently used in hemorrhagic stroke. In an ICH model, no difference in Y-maze was observed in stroke and sham-operated mice at day 3 or 10 after injury.¹⁵⁷ Similarly, stroke mice and sham controls failed to show any difference in spontaneous alternation at day 3 after SAH.¹⁵⁸ In contrast, another study found that SAH mice displayed decreased spontaneous alternation at day 3 after stroke compared to sham-operated mice.¹⁵⁹

The Y-maze utilizes natural exploring behavior of rodents to assess cognitive function. It is a simple test and requires no training. The parameter (spontaneous alternation) is easy to record, and thus this test is less demanding on the observers. The Y-maze is able to assess both short-term and long-term stroke outcomes in ischemic models, but less sensitive in hemorrhage models. Additionally, the interval should be considered if the test is performed at multiple time points. This is because the interest of exploration decreases if the animals get habituated with the maze.¹⁴⁹

3.3. Novel Object Recognition/Location Tests

The one-trial object recognition test was first developed to assess memory on rats.¹⁶⁰ This test involves a reward training using a large number of stimuli. Specifically, two identical items are placed in two arms of a Y-maze and the animal receives rewards after choosing one side. Next, a novel object is placed in the third arm and the animal gets rewards when choosing the novel arm. Then, the other two objects are replaced by two new objects, one identical to the one in the third arm and one novel. The animal is rewarded again after choosing the novel one. This trial will repeat for ten times. This test is complex, and thus a simplified version known as novel object recognition test has been developed.¹⁶¹ This test is performed in an open field instead of Y-maze. In the first session, two identical objects are placed in the back corners (Figure 1I). An animal is placed near the center of the front wall with its back toward the objects and allowed to explore for 5 minutes. In the second session, one identical object is replaced with a novel object and the animal is allowed to explore for 3 minutes (Figure 1I). The total time the animal spends exploring each object is recorded. Discrimination to the novel object indicates object recognition.

In addition, similar measurement has also been applied to assess memory associated with object location.¹⁶² In the first session, two identical objects are placed in the field in two adjacent corners (Figure 1I). The animal is placed in the field for 3 minutes and then returned to its home cage. The second session starts 15 minutes later, in which one of the objects is placed in the other adjacent corner (Figure 1I). The animal is placed in the field for 3 minutes, and the time it spends on each object is recorded. The recognition to novel object location is revealed by discrimination to the object in the new location.

The novel object recognition/location tests have been widely used in ischemic stroke. Rats with transient global ischemia spent significantly less exploration time on the novel object at day 6 after injury compared to sham controls,¹⁶³ indicating short-term visual working memory deficits after stroke. Similarly, these ischemic rats spent much less exploration time on the object at new location at day 6 after injury compared to sham controls,¹⁶³ indicating spatial memory deficits. Interestingly, long-term estradiol treatment was able to improve both visual and spatial memory in ischemic rats.¹⁶³ Another study showed that ischemic rats had spatial memory deficits at 6 but not 18 months after stroke.¹⁶⁴ Long-term visual memory deficits, on the other hand, were not detected in ischemic rats.¹⁶⁴ These results suggest that the novel object recognition/location tests are able to assess memory after ischemic stroke and evaluate neuroprotective effects of potential treatments.

The novel object recognition/location tests have also been used to assess cognitive function in hemorrhagic stroke. It was reported that both the amount and percentage of time exploring the novel object were reduced in mice with hippocampal ICH at day 21 after stroke.¹¹⁰ In addition, rats with ICH in the cortex, ventricle, and hippocampus showed lower discrimination toward the novel object at day 21 after stroke, while those with ICH in the striatum exhibited no difference compared to sham controls,¹⁶⁵ indicating that visual recognition deficits might be less correlative with striatal neuronal loss. Similarly, SAH rats displayed no discrimination toward the novel object 4 weeks after injury, while sham controls showed a significant preference to the novel object,¹⁶⁶ indicating cognitive impairment after SAH. These results suggest that the novel object recognition test is able to assess long-term cognitive function after hemorrhagic stroke.

The novel object recognition/location tests are relatively simple and objective. These tests are also very sensitive and able to assess long-term cognitive function after stroke. However, training trials are required. In addition, the result might be affected by various factors, such as the size of exploration field, shape of objects, interval between sessions, and animal activity & anxious level. For example, it was reported that the discrimination toward novel object was significantly decreased after an inter-trial interval of over 24 hours.¹⁶¹

3.4. Radial-Arm Test

The radial-arm maze was developed to study working memory on rats.^167,168 Unlike the Y-maze that primarily uses spontaneous alternation to assess memory function, the radial-arm maze uses rewards as motivation to promote decision making. The radial-arm maze has eight arms radiating from a central platform. In the original radial-arm maze test, a reward (usually food or water) is placed at the end of each arm (Figure 1J). A food-deprived animal (80-85% of normal body weight) is placed in the central platform and allowed to freely explore the maze until all rewards are found. The total number of entrances to arms before obtaining all the rewards is recorded. Low entrance number indicates that the animal is more likely to visit unfamiliar arms, suggesting a stronger working memory. It is worth noting that the original test works better in rats than in mice. For example, unlike rats, mice fail to improve the average correct choices to unvisited arms after 20 days of training, although they choose more correct arms.^167,169 Interestingly, when barriers are used to slow down the movement of mice (mice are more active than rats), their accuracy is increased.¹⁷⁰ In addition, mice show significant improvement on making correct choices in a modified six-arm maze with only three arms baited (Figure 1J).¹⁷⁰ These results indicate that a well-designed training trial is essential for memory acquisition on rodents.

The radial-arm maze is widely used in transient forebrain ischemia model, which induces hippocampal damage. In a short-term study, ischemic rats showed more visits to the unbaited arms and visited arms during days 5-11 after stroke,¹⁷¹ indicating more severely impaired short-term working memory and reference memory. In a long-term study, ischemic rats displayed increased working and reference memory errors throughout all trials up to 65 days after stroke,¹⁷² suggesting impaired long-term memory. Similar results have also been observed on mice. For example, mice with forebrain ischemia demonstrated increased working memory and reference memory errors 2 and 3 weeks after stroke, and neuroprotective treatments were able to decrease these errors.^173,174 These findings suggest that the radial-arm maze is able to assess long-term working/reference memory after ischemic stroke and evaluate neuroprotective effects of potential treatments.

In addition, the radial-arm maze has also been applied to assess cognitive function in ICH. ICH rats exhibited significantly higher number of entrances to unbaited arms 2 weeks after injury compared to sham controls, and dexmedetomidine improved their performance in the radial-arm maze.¹⁷⁵ In addition, ICH rats and sham controls showed comparable performance in the radial-arm maze 18-28 weeks after ICH,¹⁴² suggesting that this test is unable to assess long-term cognitive function after ICH. It should be noted that the radial-arm maze is rarely used in SAH.

The radial-arm maze is a relatively simple and sensitive test. It is able to assess short-term and long-term cognitive function in both ischemic stroke and ICH. Like most tests that use food as rewards, food deprivation can be used to enhance animal performance. It should be noted, however, that designing a proper trial that meets the need of the research objective is the key to a successful and efficient radial-arm maze test. Thus, preliminary study and training may be required to determine the appropriate conditions for each study.

3.5. Elevated-plus maze

Accumulating evidence suggests that patients show symptoms of anxiety after stroke.^176,177 To evaluate anxiety on animals, various behavior tests have been developed. The elevated plus-maze, a widely used anxiety test, utilizes natural behavior of rodents: they tend to stay at the corner or perimeters under fear or stress.^178,179 The elevated-plus maze consists of four arms with a “plus” shape that are 50cm above the floor. Two of the four arms are open without walls (or with 0.5cm walls to prevent mice from falling), and the other two arms are enclosed with 16cm walls (Figure 1K). An animal is placed in the enter area with its head toward the closed arm and allowed to explore for 10 minutes (Figure 1K). The number of entrances and the time it spends in each arm are recorded.¹⁸⁰ Animals with low anxiety level have higher percentage of open arm entrance and stay in the open arms longer.

The elevated-plus maze has been used to evaluate anxiety in rodents after ischemic stroke. Mixed results, however, were observed in different studies. In a global ischemia model, ischemic rats spent less time in the open arms at day 1 after injury , while significantly more time in the open arm at day 5 after injury compared to sham controls.¹⁸¹ These rats and sham controls failed to show any difference at days 15 and 30 after injury.¹⁸¹ Similarly, increased time spent in the open arms was also observed in rats at days 3 and 7 after global ischemia.¹⁸² In a transient global cerebral ischemia model, however, injured mice showed elevated anxiety level at day 2 but not day 7 after stroke, compared to sham controls.¹⁸³ In general, a bi-phase anxiety level has been commonly observed after ischemic stroke in both mice and rats: an acute increase of anxiety level in the first week and a decrease of anxiety level in the second week. In contrast to these findings, a long-term study reported that mice with MCAO spent less time in the open arms compared to sham controls 9 weeks after injury, indicating persisted anxiety after ischemic stroke.¹⁸⁴ The different results may be caused by multiple factors, including different animal strains/species, stroke methods, and training protocols.

In addition, the elevated-plus maze has also been used to assess anxiety in hemorrhagic stroke. In an ICH model, stroke animals showed no significant difference in anxiety level at day 30 after injury compared to sham controls.^49,185 In an SAH model, however, injured rats spent less time in the open arms 3 weeks after stroke.⁵⁰ Another study showed that SAH mice spent less time in the open central area at day 13 but not day 27 after surgery, although they spent similar time in open arms at both time points.⁵¹ In general, the elevated-plus maze is less effective in assessing anxiety level in hemorrhagic stroke than in ischemic stroke. This may be due to different neurological pathology in ischemic stroke and hemorrhagic stroke. Another possibility is that there are fewer studies examining anxiety in hemorrhagic stroke.

The elevated-plus maze is able to evaluate anxiety after stroke. It is an easy test, although special equipment is required. Since elevated-plus maze utilizes natural behavior of rodents to assess anxiety, training is not required for this test. It should be noted that the elevated-plus maze is more widely used in ischemic stroke than in hemorrhagic stroke.

4. Summary

Rodent models are widely used in stroke studies. Compared to other species, rodents have several advantages, including fast breeding, low maintenance cost, and higher ethical acceptance.¹¹ The neurological behaviors and higher brain functions, however, are largely different in rodents and humans. For example, unlike human patients whose neurological functions can be assessed by answering questions and following commands, rodents can only perform tasks that follow their natural responses. This makes it difficult to assess stroke outcome, especially the cognitive function and consciousness. Choosing appropriate behavioral tests that meet the objective of research is crucial in assessing stroke outcome and connecting findings in rodents to clinical trials. Here, we reviewed commonly used behavioral tests in rodent models of stroke, compare their applications in different stroke models, and discuss their advantages and disadvantages.

It should be noted that most behavioral tests were initially developed on rats for studies other than stroke, and later modified to be used in mice and/or for stroke. To obtain objective and accurate results, several factors should be considered. First, a baseline behavior should be acquired before stroke to minimize individual variation. In case that different animals show various behaviors, the baseline may be used to customize final results in these tests. In addition, trainings are required for certain behavioral tests. Since experimental animals grow up in a laboratory environment, they may lack the experience to perform certain tasks required in behavioral tests, such as removing adhesive stimuli on the limbs or keeping balance on a rotating rod. Failure to perform such tasks may be due to their lack of experience rather than functional deficits. Trainings prepare animals for such tasks and reduce individual variation, leading to more objective and consistent results.

Table 1.

Motor and sensorimotor tests in rodent models of stroke

Tests	Functions		Mice/Rats	Time Window	Advantages	Disadvantages	References
Neurological Score	Bederson score	Motor function, Body symmetry	Both	Ischemic: 24h after stroke	Easy, Simple, Effective for acute outcome evaluation	Unable to assess brain damage in certain regions, Unable to assess long-term outcomes, Low translational potential	6,8,20-22
	mNSS	Motor, Sensory, Reflex functions	Both	Ischemic: up to 28 days after stroke, ICH: up to 42 days after stroke, SAH: up to 14 days after stroke	Comprehensive evaluation, Useful for long-term stroke outcome assessment	Complex, Unable to assess brain damage in certain regions	23-31
	Garcia scale	Sensory and motor functions, Body symmetry	Both	Ischemic: up to 7 days after stroke, ICH: 24h after stroke, SAH: up to 3 days after stroke	Easy to perform, Relatively comprehensive evaluation	Limited hindlimb evaluation, Unable to assess long-term outcome	32-35
Open Field Test	Motor function, Anxiety-like behavior		Both	Ischemic: up to 6 weeks after stroke, ICH: less than 14 days after stroke, SAH: up to 27 days after stroke	Easy to perform, No training required	Results affected by protocols, environment, and habituation	36,37,39-51
Pole Test	Motor function		Mice	Ischemic: up to 20 days after stroke, ICH: 2 days after stroke	Easy, Objective, Sensitive, Able to assess long-term outcome	Cannot be used in rats, Training required	56-59
Foot-Fault Test	Motor function, Limb coordination		Both	Ischemic: up to 90 days after stroke, ICH: up to 3 weeks after stroke, SAH: up to 2 weeks after stroke	Effective, Objective, Able to assess long-term outcome	Results affected by individual variation, Pre-stroke baseline measurement required	62-71
Cylinder Test	Limb-use asymmetry		Both	Ischemic: up to 77 days after stroke ICH: up to 4 weeks after stroke	Easy, Objective, Able to assess long-term outcome	Not useful for global stroke models	73-82
Corner Test	Sensorimotor asymmetry		Both	Ischemic: up to 90 days after stroke, ICH: up to 28 days after stroke	Objective, Sensitive, Able to assess long-term outcome	Poor performance in severely injured animals or after repeated testing, Pre-stroke baseline measurement required	62,76,77,83-88
Adhesive Removal Test	Sensorimotor dysfunction, Motor asymmetry		Both	Ischemic: up to 6 weeks after stroke, ICH: up to 21 days after injury, SAH: up to 21 days after stroke	Objective, Sensitive, Able to assess long-term outcome	Training and baseline measurement required, Results affected by individual variation	4,73,92-98
Rotarod Test	Equilibrium behavior, Locomotor ability		Both	Ischemic: up to 6 weeks after stroke, ICH: up to 3 weeks after stroke, SAH: up to 3 weeks after stroke	Sensitive, Objective, Quantifiable, Able to evaluate long-term outcome	Special Equipment and training required	94,95,102-106
Wire Hanging Test	Grip strength, Endurance, Body corporation		Both	Ischemic: up to 3 weeks after stroke, ICH: up to 2 weeks after stroke, SAH: up to 2 weeks after stroke	Easy, Able to assess long-term outcome	Results affected by individual variation, Training required	71,107-110,113-120
Single Pellet Reaching Task	Skilled motor function		Both	Ischemic: up to 5 weeks after stroke, ICH: up to 32 days after stroke	Specific for skilled motor function, Able to assess long-term outcome, Able to discern compensation and recovery	Complex training requires, Results affected by individual variation and stroke severity	79,122,123,126,127,129-133

Table 2.

Cognition tests in rodent models of stroke

Tests	Functions		Mice/Rats	Time Window	Advantages	Disadvantages	References
Morris Water Maze Test	Cognition, Locomotor function		Both	Ischemic: up to 14 weeks after stroke, ICH: less than 8 weeks after stroke, SAH: up to 5 weeks after stroke	Able to assess both cognition and locomotor function, Able to evaluate long-term outcome	Long training trials required, Results affected by protocols and locomotor ability of animals	4,49,90,136-147
Y-Maze	Spontaneous alternation, Spatial memory		Both	Ischemic: up to 8 weeks after stroke, ICH: not sensitive, SAH: 3 days after stroke (controversial)	Easy to perform, Spontaneous alternation easily quantifiable, No training required	Less sensitive on hemorrhagic stroke, Results affected by animal habituation	41,155-159
Novel Object Recognition/Location Tests	Novel Object Recognition Test	Memory associated with object recognition	Both	Ischemic: 6 days after stroke, ICH: up to 21 days after stroke, SAH: up to 4 weeks after stroke	Relatively simple, Objective, Easy to perform, Able to assess long-term cognitive function	Training required, Results affected by shape of objects, animal activity, and interval between sessions	110,160-166
	Novel Object Location Test	Spatial memory		Ischemic: up to 6 months after stroke
Radial-Arm Maze	Spatial working and reference memory		Both	Ischemic: up to 65 days after stroke, ICH: up to 2 weeks after stroke	Relatively simple, Sensitive, Able to assess long-term memory function	Preliminary study and training required, Time-consuming	142,171-175
Elevated-Plus Maze	Anxiety		Both	Ischemic: up to 9 weeks after stroke (controversial), ICH: not sensitive, SAH: up to 3 weeks after stroke	Easy to perform, No training required	Special equipment needed, Results affected by individual variation and animal habituation	49-51,181-185