Measures
Platform crossings and close encounters
Quadrant time is broad. Crossings are precise.
Platform crossings count how often the animal swims over the exact spot where the platform used to be. Dr Richard Anderson's 1998 notes put it simply: a mouse with good spatial memory "will exhibit a significantly increased number of former platform position crossings" compared with the equivalent spots in the other three quadrants.
Why it matters
Crossings count how often the subject swam over the exact spot where the platform was.
More detail
Crossings are more localised than quadrant preference and work for any platform position, but they miss near misses; HVS Image therefore also counts close encounters. Whole-trial counts can favour subjects who keep searching over flexible subjects who move on quickly.
Combine crossings with heading angle, time and path to the first close encounter, and time slices to separate early and late behaviour. Maei et al. (2009), using more than 1600 mouse probe tests, found mean proximity outperformed crossings in detecting group differences.
Why it matters
Close encounters count near misses, so a subject that searches just beside the platform still gets credit.
More detail
The counter is a circle around the platform location, twice the platform diameter by default. Measures include passes, % time in the counter, and latency and path to the first close encounter, for the current, learned and every calibrated position.
Exact crossings fail to capture near-miss behaviour, so subjects that search very close to the location are not distinguished from those at chance; counter measures address this. Counter size is adjustable, so state it in your methods.
Platform crossings
Crossings count how often the subject passes over the exact platform location. Whole-trial crossings favour subjects that keep returning (perseverating) over flexible ones that check and move on, so use heading, the first close encounter and time slices as well.
More detail
HVS reports crossings of the exact location and close encounters for each calibrated position, with the learned position reported separately in probes. PASSTH00 is passes through counter 0.
The counter size is adjustable and you can re-analyse with different sizes, which helps when the exact-location count is too strict to separate groups.
Close encounters (counter: PASSTH, TIMCT%)
A close encounter is an entry into a circle around the platform position, which is less strict than an exact crossing. It shows how often, and how soon, the subject searched near the target.
More detail
Results give passes, % time in the counter, and latency and path to the first close encounter. The counter diameter setting defaults to 2.0 (close encounter radius twice the target radius).
Counter size is adjustable and you can re-analyse with different sizes. Report the build and these settings with re-analysed data.
Questions people ask
Why can platform crossings be misleading in probe trials?
Whole-trial crossings favour subjects that keep returning (perseverating) over flexible ones that check and move on. Use heading, the first close encounter, time slices and behaviour identification as well.
More detail
HVS reports crossings of the exact location and close encounters (entries into the counter, default twice the platform diameter) for each calibrated position, with the learned position reported separately in probes: passes, % time in counter, and latency and path to the first close encounter.
Going deeper
Counter size is adjustable and you can re-analyse with different sizes.
For counting crossings and near misses, see platform crossings and close passes on hvsimage.com.
Spotted a mistake, or know a better source? Tell us and we'll fix it and say so on our corrections page.