Cincinnati Water Maze (CWM) is a labyrinthine maze used to study egocentric navigation, learning, and memory. Egocentric navigation is one of the two types of local navigation described as the ability to locate places using proximal or internal cues close to the organism. The other type of navigation is called allocentric or spatial navigation that uses distal cues like visual or auditory, to locate places. Egocentric navigation involves route-based integration where an organism follows a specific set route and path integration where an organism takes a more direct route back to the starting position after exploring different locations. Egocentric navigation of particular route overtime eventually leads to the formation of an automated habit that is stored in the prefrontal cortex as a long-term memory of skilled behavior. There is an overlap between the neural networks in the brain that are involved in mediating egocentric and allocentric navigation, so a lesion or a treatment targeted to one type can cause changes in the other type of navigation.

In water mazes, water itself acts as a motivator for the subject to escape which is an advantage when compared to appetitive tasks in which motivation can differ between subjects. The reason for the difference in motivation between subjects is because, in appetitive tasks, condition or treatment under study can cause changes in body mass or produce motor deficits causing problems with palatability of the reinforcer. Water continues to act as a motivator from the first to the last trial as opposed to in appetitive tasks where motivation can decline as more rewards are gained. In most of the water maze tasks, training trials are given in a straight swim channel just to introduce the idea that escape is possible before testing in the maze. Without such trials, subjects can feel frustrated and are likely to give up searching as seen in Forced Swim Test (FST). In comparison to appetitive tasks, water maze tasks have few trail days with few limited-time trials per day as the subjects do not show off-task behaviors seen in appetitive tasks like sniffing, grooming, etc.

There are many water maze tasks for testing allocentric navigation. One of them is Morris Water Maze (MWM) developed by Morris et al., where a hidden platform submerged in a pool of water at a specific position needs to be reached by the subject during the test from different start locations. Another test used is the Radial-arm Maze (RAM) where either all the arms are baited, or some of them are baited to test spatial navigation and memory. RAM is an appetitive task that uses positive reinforcement and relies on food deprivation. A modified swimming version of RAM called Radial-arm Water Maze (RWM) was developed that uses negative reinforcement. There are a lot of different versions and protocols used for both RAM and RWM.

Cincinnati Water Maze (CWM) is one of the most compelling tests used to study egocentric learning and memory when conducted in darkness to avoid distal cues. In comparison, land-based mazes can also be used with blindfolds to study egocentric learning and memory which is not possible with water mazes



History of Cincinnati Water Maze

CWM is considered as the modified version of Beil Water Maze (BWM). Beil et al., developed BWM to study egocentric navigation. He built a 6-unit multiple T-maze with a separate straightaway maze made up of sheet metal iron placed in a glass tank. In order to remove proximal cues, the maze along with the tank was placed in a large press-board box. BWM was lightened by 9 bulbs fitted to the ceiling of the box. Two positions were marked in the maze one as the starting position ‘S’ and the other as the goal ‘G,’ where an escape platform was placed. Beil used a split litter design by pairing a male and female rat each from 14 different litters, then divided them into groups by age. Tests were conducted postnatally on day 16, 19, 22 or 29. All subjects underwent straightaway maze trials before being tested in the maze. The subjects are placed at the ‘S’ position and were allowed to escape by reaching ‘G’ position. It was observed that the younger rats made more errors by turning in the wrong direction in the first test in the maze as compared to older rats, showing an age-dependent increase in performance.

Many years later BWM was used by Polidora et al., to study effects of phenylketonuria on learning in adult rats. They changed the testing protocol by not covering the maze and increasing the number of trials from 2 to 5 per day. Later on Polidora et al., used BWM in another study where they further changed the previously used protocol by now testing the rats in the maze from point A(Start/S) to B(Goal/G) called path A, but also in reverse order, i.e., from B to A called path B. They concluded on the basis of results obtained that path A and B are not equally difficult. Butcher et al., continued work with BWM to study adult rats that were given hyperphenylalaninemic diets either prenatally or at weaning. They used 5 trials per day with 3 days testing each in path A and B. Further studies using BWM led to decreasing the length of each trial from 10 minutes to 6 minutes due to fatigue observed in subjects that reached 10 minutes mark. Further changes were made to the number of trials per day to 2 trials per day on day 1 and 3 trials on day 2 in path A followed by 2 trials per day for the next 3 days in path B. Vorhees et al., continued to use BWM to test different disease models eventually raising some concerns about the BWM. It was observed that large male rats were able to avoid swimming by propping up halfway through water using corners of the maze with the help of their legs. It was also noticed that the majority of the subjects used swim-straight-until-forced-to-turn strategy. In path A, there is 50/50 chance of turning in the correct direction as compared to path B, where the subjects will end up in a dead end of T cul-de-sac at the end of each tunnel if above said strategy is used. So path B is more difficult to learn than path A, but there were no significant differences seen and the maze was considered relatively easy. It was also observed that the long sections in the maze were received as clues by the subjects causing them to speed up to reach the end to escape.

Due to the concerns about BWM mentioned above Vorhees et al., developed Cincinnati Water Maze (CWM), a 9 unit multiple T-maze and carried out a comparative study in rats looking at effects of prenatal exposure to phenytoin. It was observed that the CWM path B was more challenging than BWM path B and showed a more significant effect of phenytoin in rats. In the next several years till date, CWM test protocols have changed as more studies were conducted with the apparatus, with the development of a more standard protocol used at present. A modified version of CWM has also been developed with a 10 unit T-maze. The subjects are placed at a starting position ‘A’ in one of the T-sections of the maze and allowed to find the escape platform at point ‘B’ in another T-section of the maze. The number of errors and latency is recorded and analyzed.

Morris Water Maze

Radial Arm Maze

Barnes Maze

Attentional Set Shifting

Apparatus and Equipment

The original design of the CWM also referred to as a multiple T-maze consisted of 9 T-sections made up of 0.95 cm thick high-density polypropylene, chemically welded to form a self-contained watertight labyrinthine maze. Each T-section has a 15 cm long stem except one stem which is 23 cm long, right and left cul-de-sacs with short walls 23 cm and back wall 61 cm long. The maze depth is 51 cm, and water is filled up to a depth of 20 cm. The maze has a 5 cm central drain opening in its flat floor connected to the floor drain by PVC pipe ending in a shut-off valve.  An overflow outlet is connected to the outer wall of the maze in one of the T-sections, 20 cm above the floor of the maze. A PVC pipe connects the overflow outlet to the pipe going into the floor drain below the shut off valve. A starting position is marked as point ‘A’ in the maze leading to a position marked point ‘B’ as the goal in different T-sections. An escape platform is hooked to the wall of the maze at point B supported by a 15 X 15 cm horizontal textured platform, 1.5 cm below the water surface.

The revised/modified CWM design has 10 T-sections. Each T-section has a 15 cm long stem, right and left cul-de-sacs with short walls 23 cm and back wall 61cm long. The modified design of CWM consists of a tank made up of high-density polyethylene. The tank walls are 1.9 cm thick and 51 cm tall with a flange of 5 cm at the top. The perimeter of the tank is surrounded by a 5 X 5 cm stainless steel reinforcing bar mounted just beneath the flange. The tank floor is sloped over 10 cm distance from the edge to the center of the tank, where it opens up into a 5 cm drain opening. The drain opening is connected to the floor drain by a PVC pipe ending with a shut-off valve. An overflow outlet is located in the tank side wall 25 cm above the tank floor. The maze is separate from the tank and sits inside it. The maze walls are made up of 0.95 cm thick PVC and welded together chemically. The maze is welded in six sections with stainless steel screws joining sections. The sections are fixed to the underlying PVC subfloor. The PVC subfloor has equally spaced 1.9 cm diameter holes that allow water and waste to flow through the sloping tank floor and eventually draining out from the central opening.

A separate straight water channel is used for training of test subjects before maze testing. The channel is 244 cm long, 15 cm wide and 51 cm deep. The channel is filled with water to a depth of 20 cm. A submerged platform is located at one end of the channel.

Training Protocol

Clean the entire apparatus to remove any unwanted cues that can influence the performance of the subjects. The tracking and recording of the trials can be performed using tracking and video system such as the Noldus EthoVision XT. Before testing subjects in CWM, subjects are given 4 consecutive trials in the straight water channel under standard lighting. The subject is placed at one end of the channel facing the end wall and allowed 2 minutes to find the platform.

Fill up the maze with water a day before the testing so that the temperature can equilibrate with the room temperature of around 20-230C. Monitor the water temperature daily and change the water every 1-2 days. After a day or two of straight water channel training, maze test is conducted in a different room. The subject is placed at the start position ‘A’ and allowed 5 minutes to explore and find the escape platform at ‘B.’ The subjects are given 2 trials per day. At the end of 5 minutes in trial 1, if the subject is able to escape, immediately give trial 2. If the subject fails to escape at the end of 5 minutes in trial 1, remove it and place it in the holding cage for 5 minutes before giving trial 2. Another protocol for maze testing is to give trial 1 to a group of subjects in succession followed by trial 2 in the same order.

The maze test can be conducted under standard lighting or infrared lighting. If the test is conducted under infrared lighting, the room door is tightly sealed with weather stripping to prevent any amount of light entering the room. Infrared LED bulbs are mounted around the maze, and a light-sensitive video camera is mounted above the maze to observe the movement of the subject with the monitor placed in an adjoining room. Subjects are placed in the dark room away from the maze for 5 minutes for adaptation. Place each subject at the start position ‘A’ in the maze and exit the room as quickly as possible closing the door tightly behind. The subjects are tested for 15-20 days under infrared lighting and 5-6 days under standard lighting with 2 trials per day for both settings.

Evaluation of the effect of phenylethylamine drugs on egocentric learning in the Cincinnati Water Maze

Vorhees et al., used (+)-Methamphetamine (MA), (+/-)-3, 4-Methylenedioxymethamphetamine (MDMA), (+)-amphetamine (AMPH), and (+/-)-fenfluramine (FEN) in rats to study effects on egocentric learning. All four drugs have CNS effects, and at higher doses, each causes a prolonged reduction in brain dopamine (DA) or serotonin levels. Adult male Sprague-Dawley CD IGS rats were given drugs subcutaneously in 4 doses at 2-hour intervals. On day 13 after treatment, the rats were given 4 trials with 2 minutes per trial in straight swimming channel. On day 14 the subjects were tested in CWM under LED infrared emitters for 21 days. Two trials were given per day with a 5-minute time limit for each trail and an intertrial interval of 5 minutes. Latency to escape and the number of T, stem, and start return errors were recorded separately on each trial. It was found that there were no significant changes in learning with MDMA and FEN in comparison to AMPH and MA that showed a significant increase in errors and latency. It is known that FEN and MDMA affect serotonin and AMPH and MA affect dopamine in the CNS. Hence the data suggests that dopamine may mediate egocentric learning. The results of this study can be useful in the future to further understand the damages to learning and memory mechanisms caused by drugs that can cross the blood-brain barrier.

Evaluation of Targeted Mutations in the Na, K-ATPase Alpha 2 Isoform Confer Ouabain Resistance and Result in Abnormal Behavior in Mice

Schaefer et al., conducted one of the first experiments in mice using CWM to determine if the ouabain binding site of the α2 isoform of Sodium and potassium-activated adenosine triphosphatases (Na, K-ATPase) plays a physiological role in CNS function. Ouabain like compounds functions as stress-related hormones that are endogenously released in many mammals.  Mice were genetically modified to make ouabain-sensitive α2 isoform resistant (α2R/R) without affecting the basal Na, K-ATPase enzymatic function. Before testing subjects in CWM under infrared lighting, they were trained in smaller MWM. The subjects were tested for 15 days in CWM with two trials given per day with a 5-minute time limit for each trail and an intertrial interval of 15 minutes. The number of errors and latency were recorded. It was found that mice with α2R/R showed a significantly higher number of errors and increased latencies in the maze as compared to wild-type mice. Hence, CWM task can be useful for testing genetic manipulations in mice in which disruptions of egocentric substrates are suspected in the regions where egocentric neuronal network are found.

The role of Dopamine depletion in egocentric Cincinnati water maze performance and allocentric Morris water maze learning.

Braun et al., conducted a study in rats to distinguish between regions in neostriatum that is implicated in egocentric and spatial learning. The study used CWM to test egocentric navigation and MWM to test spatial learning. It was thought that dopamine in dorsolateral striatum (DLS) is implicated in egocentric learning while dopamine in the dorsomedial striatum (DMS) is implicated in spatial learning. Adult male Sprague–Dawley CD IGS rats were acclimatized for 1 week before undergoing surgery, where 6-hydroxydopamine (6-OHDA) was injected in DLS or DMS. The subjects were allowed to recover for 2 weeks before being tested in straight water channel to assess swimming abilities for 4 consecutive trials with a trial limit of 2 minutes. The day after straight channel trials the subjects were tested in CWM under infrared lighting with two trials per day with a 5-minute time limit for each trail and an intertrial interval of 5 minutes. The subjects were then tested in MWM the next day. It was observed that the DA levels decreased by 75% in DLS after 6-OHDA administration causing deficits in egocentric navigation in CMW but not in allocentric learning in MWM task. Similarly, DA levels decreased by 62% in DMS causing deficits in egocentric navigation in CMW but not in allocentric learning in MWM task. The results of the study show for the first time the role of DA in both DLS and DMS in neostriatum in egocentric rather than spatial navigation and memory.

Data Analysis

The data recorded during the maze test is as follows

  • Number of errors
  • Latency

Errors are defined by an imaginary line at the base of each T-section and the end of each arm of T. An error is made when the subject crosses these imaginary lines when entering stem, or either arm of the T, or returning back to the start arm after leaving it.

Strengths & Limitations


CWM can be used to test for both egocentric and spatial navigation with either infrared or standard lighting respectively. CWM has a more complex pathway to find the escape platform hence it is considered as a better test than other water mazes. CWM like other water mazes requires very little training of test subjects especially when testing rodents as they are natural swimmers and require very few trials to learn that escape is possible with searching. No additional reward or reinforcement is needed for the subjects to complete the task as water itself acts as a motivator to escape.


CWM can be challenging to use with certain animals like mice due to their species-specific characteristics of floating in water rather than actively searching and they tend to search along the edges (Thigmotaxis). Mice also have a large surface area to volume ratio and tend to lose more heat, leading to hypothermia which will slow down search activity and show fatigue. CWM can cause thermal stress to subjects from water temperature. Warm water tends to slow down learning performance, so the water temperature is maintained at room temperature 20-220C which promotes motivation and prevents stress (Vorhees et al., 2016). CWM has not been yet tested in smaller mammals other than rats and mice.

Summary and Key Points

  • Cincinnati Water Maze (CWM) is a labyrinthine maze used to study egocentric navigation, learning, and memory.
  • CWM is modified version of BWM made up of 9 unit multiple T-maze with a start location at point A and escape platform located at point B.
  • The subjects undergo straight water channel training before being tested in the CWM.
  • The test trials are conducted under infrared lighting to study egocentric navigation as it removes distal cues
  • Number of errors and latency to escape is recorded.
  • CWM is used to test effects of different drugs and lesions in disease models and genetically modified animals on egocentric learning, navigation, and memory.


Biel W.C (1940). Early age differences in maze performance in the albino rat. J. Genet. Psychol. 56(2), 439–453. DOI: 10.1080/08856559.1940.10534511

Polidora V.J, Boggs D.E, Waisman H.A (1963). A behavioral deficit associated with phenylketonuria in rats. Proc. Soc. Exp. Biol. Med. 113, 817–820. DOI: 10.3181/00379727-113-28500

Polidora V.J, Cunningham R.F, Waisman H.A (1966a). Phenylketonuria in rats: reversibility

of behavioral deficit. Science 151(3707), 219–221. DOI: 10.1126/science.151.3707.219

Butcher, R.E., Vorhees, C.V., Berry, H.K., 1970. A learning impairment associated with induced

Phenylketonuria. Life Sci. 9, 1261–1268. DOI: 10.1016/0024-3205(70)90266-3

Morris R (1984). Developments of a water-maze procedure for studying spatial learning in the rat. J. Neurosci. Methods 11(1), 47–60. DOI: 10.1016/0165-0270(84)90007-4

Vorhees C.V (1987). Fetal hydantoin syndrome in rats: dose-effect relationships of prenatal phenytoin on postnatal development and behavior. Teratology 35(3), 287-303. DOI: 10.1002/tera.1420350302

Vorhees C.V (1987b). Maze learning in rats: a comparison of performance in two water mazes in progeny prenatally exposed to different doses of phenytoin. Neurotoxicol. Teratol. 9(3), 235–241. DOI: 10.1016/0892-0362(87)90008-0

Vorhees C.V, Weisenburger W.P, Acuff-Smith K.D, Minck D.R (1991a). An analysis of factors influencing complex water maze learning in rats: effects of task complexity, path order and escape assistance on performance following prenatal exposure to phenytoin. Neurotoxicol. Teratol. 13(2), 213–222. DOI: 10.1016/0892-0362(91)90013-M

Vorhees CV, Acuff-Smith KD, Weisenburger WP, Minck DR, Berry HK (1992). Branched chain amino acids improve radial-arm maze acquisition and water maze forced-choice learning in rat offspring exposed in utero to hyperphenylalaninemia. Neurotoxicol Teratol. 14(1), 35-41. DOI: 10.1016/0892-0362(92)90026-7

Schaefer T.L, Lingrel J.B, Moseley A.E, Vorhees C.V, Williams M.T (2011). Targeted mutations in the Na, K-ATPase alpha 2 isoform confer ouabain resistance and result in abnormal behavior in mice. Synapse 65(6), 520–531. DOI: 10.1002/syn.20870

Vorhees CV, He E, Skelton MR, Graham DL, Schaefer TL, Grace CE, Braun AA, Amos-Kroohs R, Williams MT (2011). Comparison of (+)-methamphetamine, ±-methylenedioxymethamphetamine (MDMA), (+)-amphetamine and ±-fenfluramine in rats on egocentric learning in the Cincinnati water maze. Synapse 65(5), 368–378. DOI: 10.1002/syn.20854

Braun AA, Amos-Kroohs RM, Gutierrez A, Lundgren KH, Seroogy KB, Skelton MR, Vorhees CV, Williams MT (2015). Dopamine depletion in either the dorsomedial or dorsolateral striatum impairs egocentric Cincinnati water maze performance while sparing allocentric Morris water maze learning. Neurobiol Learn Mem.118, 55-63. DOI: 10.1016/j.nlm.2014.10.009

Vorhees CV, Williams MT (2016).  Cincinnati water maze: A review of the development, methods, and evidence as a test of egocentric learning and memory. Neurotoxicol Teratol 57, 1-19. DOI: 10.1016/