Morris Water Maze Protocol: IACUC Requirements and Welfare Limits

The Morris water maze (MWM) is one of the most widely used behavioral assays in rodent neuroscience, testing hippocampal-dependent spatial learning and memory by requiring an animal to swim to a submerged, hidden platform using only distal spatial cues. Because it is a forced-swim paradigm run over repeated daily trials, it carries genuine welfare stakes that an Institutional Animal Care and Use Committee (IACUC) protocol must address explicitly, not as an afterthought to the experimental design. This guide covers what a protocol needs to specify, the confounds that most often undermine the test’s validity, and the blinding and randomization practices ARRIVE 2.0 expects reviewers and readers to see documented.

What the Test Measures, and Why Welfare and Validity Are Linked

In the standard task, an animal is released into a circular pool of opacified water from one of several start locations and must locate a fixed, submerged escape platform using room-based visual cues rather than intra-maze cues (which are deliberately absent or symmetrical). Learning is typically indexed by escape latency and path length across repeated trials, with a probe trial (platform removed) used to assess memory of the platform’s former location via time spent in the target quadrant. The task’s sensitivity to hippocampal and cortical dysfunction has made it a mainstay of models ranging from aging and Alzheimer’s disease to traumatic brain injury and developmental toxicant exposure.

The compliance point that follows from this design is straightforward: MWM data quality and animal welfare are not competing concerns. A cold pool, an exhausted swimmer, or a poorly controlled cue set produces both a distressed animal and a noisy, harder-to-interpret dataset. An IACUC protocol that specifies the parameters below is protecting the animal and the experiment at the same time.

What an IACUC Protocol Must Specify

Scientific Justification and the 3Rs

Every MWM protocol needs an explicit 3Rs justification. Replacement is rarely available for this specific endpoint — no validated in vitro or computational assay yet substitutes for whole-animal spatial memory testing, though investigators should document that they considered non-animal alternatives, consistent with New Approach Methodologies (NAMs) guidance where a genuine option exists. Reduction is served by a well-powered design (an a priori sample-size calculation, not an inherited group size from a prior paper) and by protocol features that avoid discarding data to unreliable measures — see the confounds section below. Refinement is where most of the protocol’s welfare content lives: water temperature, trial limits, humane removal criteria, and post-trial thermal recovery, each covered next.

Water Temperature and Animal Welfare Limits

Water temperature is a primary welfare and validity variable, not a housekeeping detail. The methodology most commonly cited for standardizing MWM procedure, Vorhees and Williams’ 2006 Nature Protocols paper, specifies water equilibrated to ambient room temperature — approximately 19–22°C for rats — as sufficient for good performance without inducing fatigue or hypothermic stress; for mice, the same protocol recommends water in the 20–22°C range, noting that colder water increases the incidence of floating (a non-swimming, non-learning behavior discussed below). A protocol should state the target temperature range explicitly, how it will be verified and logged for each session (not assumed from a thermostat setting), and the corrective action if water drifts outside range. Water that is too cold risks hypothermia and increased floating; water that is too warm removes the mild aversive stimulus the task depends on and can independently affect swim speed.

Trial Number, Trial Duration, and Session Limits

The protocol should specify trials per day, maximum trial duration, and inter-trial interval, all of which bound cumulative swim exposure per animal per day. The Vorhees and Williams protocol describes a common structure of four trials per day with a per-trial ceiling — commonly 60 seconds for mice and up to 120 seconds for rats — beyond which an animal that has not located the platform is guided or placed onto it rather than left to continue swimming, with inter-trial intervals (on the order of 15 seconds, longer for mice) sufficient to prevent excessive fatigue between trials. A protocol that leaves trial duration or daily trial count unbounded, or that does not specify what happens when an animal fails to find the platform within the ceiling, is missing a welfare control an IACUC reviewer should flag.

Humane Endpoints for Non-Swimmers

Not every animal swims productively. Some fail to swim at all (persistent floating) or swim but never engage in goal-directed search, and this is a documented, expected occurrence rather than a rare failure. A compliant protocol specifies a concrete rule for this population: how many consecutive non-swimming trials trigger removal from the pool for that session, at what point an animal is excluded from the study entirely rather than repeatedly re-tested, and — critically — that floating frequency is tracked and statistically compared across groups rather than simply discarded, since a treatment effect that increases floating is itself a result, not noise to be edited out. Persistent floating, distress signs (repeated diving, prolonged immobility with no righting), or physical signs of hypothermia are the triggers a protocol should name for removing an animal from further testing and, where indicated by the attending veterinarian, from the study.

Drying and Thermal Recovery After Each Session

Because rodents lose heat rapidly in water, post-trial thermal recovery is a required, not optional, protocol element: animals are typically dried with a towel and placed under a warming lamp or on a heated pad, monitored until fully dry and normothermic, before being returned to the home cage. The protocol should state how recovery is confirmed (visual check, time-based minimum, or both) and who is responsible for it, and this applies to every animal on every testing day, not only ones that showed obvious distress during the trial.

Personnel Qualification

Handlers running MWM sessions need documented training in rodent handling, recognition of the humane-endpoint signs above, and the specific pool procedures for that protocol — the same personnel-competency expectation the Guide for the Care and Use of Laboratory Animals applies across all animal procedures, not a lower bar because the task is behavioral rather than surgical.

Confounds That Threaten Validity

Beyond welfare, several design errors are common enough in MWM literature that an IACUC reviewer and the investigator should both be checking for them before data collection starts.

Visual (Distal) Cue Placement

The task depends on stable, sufficiently distinct room cues visible from the pool and unrelated to the platform’s own position. Cues that move between sessions, that are only visible from some start positions, or that inadvertently correlate with the platform location (e.g., a cue closer to the platform than to other equidistant points) can let an animal solve the task through a strategy other than the spatial mapping the test is meant to measure — and the same defect degrades welfare, since a poorly solvable task means more prolonged swimming for the animal. Cue configuration should be fixed for the duration of acquisition testing and documented in the protocol and eventual methods write-up.

Thigmotaxis

Thigmotaxis — wall-hugging swim behavior — is a well-recognized alternative strategy that keeps latency and path-length measures low without reflecting spatial learning; animals showing high thigmotaxis are essentially failing to engage the task rather than succeeding at it efficiently. Excessive thigmotaxis can reflect anxiety-like behavior, visual impairment, or simply an animal that has not yet learned to search, and current MWM tracking software typically scores time-in-annulus or a comparable thigmotaxis index precisely so this can be checked before latency data is trusted as a learning measure.

Swim-Speed Differences Confounding Latency

Escape latency is a compound measure of both navigation accuracy and swim speed, and any manipulation, genotype, or drug that changes swim speed independently of spatial ability will distort latency-based conclusions. A treatment that slows animals will look like an impaired learner on latency alone even with an intact spatial strategy; a treatment that speeds animals up can mask a real deficit. The standard mitigation is to report path length (distance swum to reach the platform) alongside or instead of latency, since path length is far less sensitive to swim-speed differences, and to record and report swim speed itself so readers can assess whether it differed between groups. A protocol or methods section that reports only latency, with no speed data and no path-length cross-check, has not ruled out this confound.

Blinding and Randomization per ARRIVE 2.0

The ARRIVE 2.0 Essential 10 lists randomization and blinding/masking among the minimum items every animal study should report, and MWM studies are a common place for both to go undocumented. Randomization should cover both group allocation (which animals receive which treatment, not experimenter convenience of cage or arrival order) and, within a session, the sequence of start positions used across trials and animals, since a fixed start-position sequence can itself become a learnable, non-spatial shortcut. Blinding should extend to whoever scores latency, path length, and thigmotaxis from the tracking software — group identity should not be visible to the scorer during video analysis, and where fully automated tracking is used, the protocol should state that the analysis parameters (platform zone definition, thigmotaxis annulus width) were set identically and blind to group before scoring began. See the full ARRIVE 2.0 checklist entry for how these items map onto the rest of a methods write-up.

Documenting the Protocol for IACUC Review

Pulling the above together, a submission-ready MWM protocol section should state: the scientific justification and 3Rs analysis; pool dimensions, water temperature range and how it is verified per session; number of trials per day, per-trial time limit, and inter-trial interval; the exact criterion for classifying and removing a non-swimming or distressed animal, and whether that animal is excluded from analysis or from the study; the post-trial drying/warming procedure and how recovery is confirmed; personnel named and their training record; and the randomization and blinding scheme for group allocation, start-position sequencing, and outcome scoring. Institutions typically require this level of specificity under PHS Policy and, where the species covered are within its scope, the Animal Welfare Act — the IACUC itself is the body that reviews this level of protocol detail rather than a general procedure description, per its function described in the research animal oversight framework. A behavioral assay with similarly explicit welfare and confound controls to compare against is the rotarod motor test, which shares the same IACUC review logic for a non-cognitive endpoint.

Frequently Asked Questions

Does the Morris water maze count as a survival or non-survival procedure for IACUC classification?

It is a survival, non-terminal behavioral procedure in the overwhelming majority of protocols — animals are returned to the home cage after drying and thermal recovery. IACUC pain/distress categorization still applies and should reflect the repeated forced-swim exposure, even though the procedure itself does not involve surgery, injection, or euthanasia as an endpoint of the test itself.

What water temperature range should a protocol specify?

Published methodology commonly cites approximately 19–22°C for rats and 20–22°C for mice, equilibrated to ambient room temperature, as sufficient for reliable performance without cold stress. The exact range should be justified for the species and strain used and verified each session, not assumed.

How should a protocol handle an animal that will not swim?

Specify a concrete removal criterion (e.g., a set number of consecutive non-swimming trials), what happens to that animal for the remainder of the session and study, and a requirement to track and statistically report floating frequency by group rather than silently excluding the data.

Why does escape latency alone not settle whether an animal learned the task?

Latency conflates spatial accuracy with swim speed, so a genotype or treatment that changes speed can produce a latency difference unrelated to learning. Reporting path length and swim speed alongside latency, and checking thigmotaxis, lets reviewers and readers distinguish a real spatial deficit from a motor or motivational one.

Is the Morris water maze required to be blinded?

ARRIVE 2.0 lists blinding/masking as an Essential 10 reporting item for animal studies generally, and it applies directly here: the person allocating animals to groups and the person scoring escape latency, path length, and probe-trial performance should not have visible access to group identity while doing so.

Primary Sources

• Vorhees, C.V. & Williams, M.T. “Morris water maze: procedures for assessing spatial and related forms of learning and memory.” Nature Protocols 1(2), 848–858 (2006).
• NC3Rs / ARRIVE 2.0 guidelines (Percie du Sert et al., 2020) — Essential 10 reporting items.
• National Research Council. Guide for the Care and Use of Laboratory Animals, 8th edition.