Working Memory3–5 minBurden: MediumEMA: Low

Route Memory

A spatial span task with routes: a marker walks a route over a grid map with landmarks, one neighbouring square at a time, and the participant taps the route back in order. Routes grow by one square until both routes of a length are wrong. Every map and route is generated from a seed.

Visuospatial short-term memoryRoute memorySpatial span
Category
Working Memory
Typical duration
3–5 min
Participant burden
Medium
EMA suitability
Low

Try this task

Configure parameters and run an interactive preview exactly as participants will experience it. No data is recorded.

Configure preview

Adjust parameters below, then start the preview on the right.

Include practice trials

Shown with feedback before the main task

Task parameters

Default 2, as in the standard Corsi procedure (Kessels et al., 2000).

Default 9 (Kessels et al., 2000). At most half the squares of the map (12 on this map).

Default 2 (Kessels et al., 2000). The next length follows when at least one route of a length was right; the task stops when all were wrong.

Default 5 × 5 (our choice; no published grid size).

Default 500 ms, as each block is lit in eCorsi (Brunetti et al., 2014).

Default 500 ms: one square per second, the Corsi rate (Kessels et al., 2000; Brunetti et al., 2014).

Drawn icons (tree, fountain, tower, house, flag, bridge) as reference points; default 4 (our choice).

Default none (our choice; the Corsi procedures set none). A route not finished in time counts as an omission and as wrong.

With feedback on every tap. Default 3 (eCorsi: three practice sequences, Brunetti et al., 2014).

Default 3 squares (eCorsi, Brunetti et al., 2014).

Empty: each participant gets a new map and new routes, and the seed they were drawn from is stored with the result. The same seed gives every participant the same map and routes.

The span procedure of the Corsi block-tapping test (Kessels et al., 2000) applied to routes on a map, as the Walking Corsi Test applies it to paths walked on the floor (Piccardi et al., 2008). Every step goes to a neighbouring square and a route never crosses itself, so routes are likely easier to hold than Corsi sequences of the same length (path configuration matters: Busch et al., 2005; Parmentier et al., 2005); expect longer spans than on the Corsi test and check for a ceiling at the longest route. About 3 min with these settings (assuming a span of about 6). Outputs: span, routes tapped correctly, total score (span × correct routes), squares tapped in the right place, median time to the first tap and between taps; each trial stores the map, the route, the taps and the time of every tap.

3–5 minBurden: MediumEMA: Low

This is a researcher preview. No participant data is recorded.

Simulated participant view

9:41

Route Memory

A spatial span task with routes: a marker walks a route over a grid map with landmarks, one neighbouring square at a time, and the participant taps the route back in order. Routes grow by one square until both routes of a length are wrong. Every map and route is generated from a seed.

No data is recorded

Participant experience on smartphone

A square map (5 × 5 by default) fills the width of the screen, with a few landmarks drawn as dark icons (tree, fountain, tower, house, flag, bridge) told apart by shape, not colour. The marker lights up one square at a time; then every square is a button. A tapped square lights up briefly; in practice it turns green or red, and the route is shown numbered afterwards.

When to use

For a spatial span measure with routes instead of scattered locations, as in the route and topographical memory literature (Walking Corsi Test), and as practice for the route tasks of selection tests. The span procedure is the standard Corsi procedure (Kessels et al., 2000), so span and total score are defined as there. Maps and routes are generated, so sessions can use new routes (no seed) or one fixed form (a set seed).

When not to use

Not a normed test and not a substitute for the Corsi test: routes along neighbouring squares never cross themselves and are likely easier to hold than Corsi sequences of the same length, so spans are not comparable with Corsi norms. Not a quick EMA probe: the span procedure takes a few minutes and its length depends on performance. Not a navigation test in large-scale space (the map is seen from above).

How to use in a study

Keep the defaults for the standard procedure: routes of 2 to 9 squares, two per length, the next length when at least one route of a length was right, stop when both were wrong; 500 ms per square and 500 ms between squares (one square per second); three practice routes of three squares with feedback. Leave the seed empty for a new map and routes per participant and session, or set one to give everyone the same form. If many participants reach the longest route, raise it (up to half the squares of the map) or use a larger map.

Researcher-configurable parameters

  • Shortest and longest route: default 2 and 9 squares (Kessels et al., 2000); the longest at most half the squares of the map
  • Routes per length: default 2 (Kessels et al., 2000); the task stops when all routes of a length are wrong
  • Time on each square and pause between squares: default 500 ms each, one square per second (eCorsi, Brunetti et al., 2014)
  • Map size: default 5 × 5 squares (our choice), 4 × 4 to 7 × 7
  • Landmarks: default 4 (our choice), up to 6, spread over the map; the map stays the same for the whole run
  • Time limit per route: default none (our choice); a route not finished in time counts as an omission
  • Practice: default 3 routes of 3 squares, every tap marked right or wrong (eCorsi, Brunetti et al., 2014)
  • Seed: empty (a new map and routes per run; the seed is stored with the result) or a number (the same for everyone)

Outputs collected

  • Span: the longest route tapped correctly (0 if none)
  • Routes tapped correctly, and routes shown
  • Total score: span × routes tapped correctly (Kessels et al., 2000)
  • Squares tapped in the right place, over all routes (partial credit)
  • Median time from the end of the route to the first tap, and between taps, on correct routes
  • Every trial: the map and its landmarks, the route (squares in order), its length and turns, the taps, the time of every tap from the end of the route, squares in the right place, correct, time-outs
  • The seed, from which the map and every route of the run can be rebuilt

Interpretation notes

The span procedure is the Corsi procedure of Kessels et al. (2000): both routes of a length are always given, the next length follows when at least one was right, and the run stops when both are wrong or after the longest length. Span is the longest length tapped correctly; the total score (span × correct routes) also counts the routes tapped correctly on the way. A route is right only when every square is tapped in order; squares in the right place give partial credit, as in partial-credit scoring of span tasks (Conway et al., 2005). Unlike Corsi sequences, every step goes to a neighbouring square (no diagonal steps), no route crosses itself, and no more than two steps in a row go the same way, so a route cannot be a single straight line. Path length, crossings and angles change how hard spatial sequences are (Busch et al., 2005; Parmentier et al., 2005): expect longer spans than on the Corsi test, and check for a ceiling at the longest route. The map and its landmarks stay the same for the whole run, as the Corsi board does. Times are taken at touch-down from the moment the marker has left the last square; this timing has not been validated on real devices. The task has no norms; compare groups or sessions within a study.

Scientific evidence

  • The standardised Corsi procedure gives two sequences per length, from 2 to 9 blocks, and stops when both sequences of a length are wrong; span and total score were normed in 70 healthy adults and 70 patients with cerebral lesions, with right-hemisphere patients doing worse (Kessels et al., 2000).
  • On a tablet (eCorsi), with each block lit for 500 ms at one block per second, young adults' forward span was about 6.1 and older adults' about 4.8 (Brunetti et al., 2014).
  • The Walking Corsi Test applies the same procedure to paths walked over a floor layout; it detects topographical memory deficits in people without other memory impairments and was standardised in 289 people aged 15 to 86 (Piccardi et al., 2008, 2013).
  • From about 5.6 years on, children's spans are larger on the tabletop Corsi test than on the Walking Corsi Test (Piccardi et al., 2014).
  • How well a spatial sequence is recalled depends on its path: its length, crossings and angles (Busch et al., 2005; Parmentier et al., 2005).