Measures
Search strategies
"Rats can cheat!" Strategy analysis is how you catch them.
The original notes warned that instead of learning where the platform is, animals "can learn search strategies", such as swimming at a set distance from the side. Classifying each swim by strategy shows how an animal solved the task, and how that changes with training.
Why it matters
Chaining is circling at the platform's distance from the wall, a way to find it without a spatial map.
More detail
Chaining reflects a non-spatial, egocentric strategy and can occur even when latency looks good. In probe trials, chaining animals keep the same radius; target-quadrant time stays near chance and crossings are incidental.
After full training chaining typically indicates impaired spatial learning, but it may instead mean strong spatial cues are lacking, or that learning is at an early stage. Tracking circling and chaining across trials maps the transition from non-spatial to spatial strategies; persistent chaining is a reason to review distal cues.
Pool circling and chaining
Chaining is circling at a fixed distance from the wall, roughly the platform's distance: a non-spatial, procedural strategy that can coexist with good latencies.
More detail
In probes with chaining, target-quadrant time stays near chance and crossings are incidental. Pool circling counts complete loops enclosing the pool centre; doubling back does not affect the count. Early in learning chaining is common.
Chaining in a probe after full training suggests spatial learning is impaired or absent, or spatial cues are lacking. Track circling and chaining across trials to map the move to spatial search; if chaining persists, review the set-up (distal cues).
Questions people ask
How does HVS identify search strategies and behaviours automatically?
With predefined, transparent quantitative rules applied to time-stamped position data, not machine learning or path images. 1) Behaviours include direct paths, initial spatial intent, target-focused search, target scanning, chaining, broad coverage, thigmotaxis and inactivity. 2) Select 'Main behaviors' in export to list them. 3) Upgrading applies it to existing trials too.
More detail
A search strategy is the overall way a subject looks for the goal, for example heading straight to it, scanning near it, chaining round the pool at the platform's distance or hugging the wall; HVS Image identifies these with predefined, transparent rules applied to the time-stamped positions. The rules capture hesitation, stopping and orientation phases and were developed with human VR and rodent data. Behaviours are evaluated relative to the current target, the learned location in probes, previous targets and other calibrated locations. They appear on the path plot, the on-screen summary and the .csv export (one row per trial with measures and labels). Allocentric/egocentric identification with moved proximal cues is an add-on.
Going deeper
Because the rules are explicit thresholds rather than a trained classifier, the same trial always receives the same labels and they can be reported in a methods section. Cite the Direct Path technical definition doi:10.5281/zenodo.21870806 and the inactivity definition doi:10.5281/zenodo.21797205. The automated rules are recent and have no peer-reviewed validation papers yet.
How do I get pool circling to tell circling from other search patterns?
Tick Pool Circling under 'Advanced' on the analysis screen, then re-make the SUM and CST. 1) Per trial: double-click, 'OK', 'Full Analysis' shows pool-circlings, clockwise and anticlockwise. 2) The ratio of circles to path length separates circling from zig-zag or quadrant searching.
More detail
SUM lines are 'Circling', 'Clckwise', 'Anticlck' (names vary by version). Include them via 'Export Results / Setup Custom' > '2. Define' on a SUM made with circling ticked. Pool circling counts complete loops enclosing the pool centre; doubling back does not affect the count.
Going deeper
Chaining (circling at about the platform's distance from the wall) is now identified automatically along with Direct Path classes.
How should I interpret chaining?
Chaining is circling at a fixed distance from the wall, roughly the platform's distance, a non-spatial, procedural strategy that can coexist with good latencies. Chaining in a probe after full training suggests spatial learning is impaired or absent, or spatial cues are lacking.
More detail
In probes with chaining, target-quadrant time stays near chance and crossings are incidental. Track circling and chaining across trials to map the move to spatial search; early in learning chaining is common.
Going deeper
If chaining persists, review the set-up (distal cues).
HVS Image explains circling and chaining, one of the strategies described here.
Sources
- Garthe et al. (2009). Adult-Generated Hippocampal Neurons Allow the Flexible Use of Spatially Precise Learning Strategies. PLoS ONE 4(5):e5464. doi:10.1371/journal.pone.0005464
Spotted a mistake, or know a better source? Tell us and we'll fix it and say so on our corrections page.