## PhenoTyper: Initial discrimination and reversal learning (CognitionWall)

**LSID: http://syli.cz/urn:lsid:public.sylics.com:automatedtest:9E48-295G-276A**

**Materials and Methods**

Mice were singly housed on sawdust in standard Makrolon type II cages enriched with cardboard nesting material for at least one week prior to experiments, with water and food ad libitum (7:00/19:00 lights on/off; providing an abrupt phase transition).

Activity in the home cage was automatically recorded by video tracking in specially designed cages (PhenoTyper model 3000, Noldus Information Technology, www.noldus.com/phenotyper). Each cage contains a top unit with built-in hardware for video-tracking, that is, an infrared-sensitive video camera. The latter provide constant and even illumination of the cage. An infrared filter placed in front of the camera prevents interference with room illumination. This method allows continuous behavioural recordings in both dark and light periods. EthoVision was used as video tracking and trial control software. PhenoTyper cages were connected in a specially designed computer network. The cages (L =30 × W =30 × H =35 cm) were made of transparent Perspex walls with an opaque Perspex floor covered with bedding based on cellulose. A feeding station and a water bottle were attached on to two adjacent walls outside of the cage. A shelter (height: 10 cm, diameter: 9 cm; non-transparent material) was fixed in one of the corners. The X-Y coordinates of the centre of gravity of mice were acquired and smoothed using EthoVision software and processed to generate behavioural parameters using AHCODA analysis software (Synaptologics BV, Amsterdam, The Netherlands, http://www.sylics.com/bioinformatics/ahcodatm-data-analysis/).

**Associated Mammalian Phenotype Ontologies (MGI)**

- MP:0001362 abnormal anxiety-related response

- MP:0002062 abnormal associative learning

- MP:0002063 abnormal learning/memory/conditioning

- MP:0001392 abnormal locomotor behaviour

**Initial Discrimination learning (1 or 2 days)**

The CognitionWall is a wall with three entrances that is placed in front of a food dispenser in the automated home-cage. During the CognitionWall initial discrimination learning task, mice have to learn to enter the CognitionWall through the left entrance to obtain food (every fifth entrance through left is rewarded with one food pellet). The primary outcome measure to assess learning in this task is the number of entries through the CognitionWall before mice reached the learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). The

__number of entries to reach this 80% criterion__is plotted in a survival plot, with

__log rank statistics__, in order to be able to include mice that did not reach the learning criterion during the task.

The primary outcome measure to investigate general activity differences is the

__total number of entries__through any of the holes during the initial discrimination learning phase. Individual mice that make very few entries, and therefore do not experience the operant schedule to the extent that produces learning, should not be included in the evaluation of initial discrimination learning.

**Reversal learning (2 days, following 2 days of initial discrimination learning)**

Following 2 days of initial discrimination learning, using another entrance in the same CognitionWall, an automated reversal learning test is implemented. Mice are rewarded with a food reward when they choose to pass through the right entrance. The primary outcome measure to assess learning in this task is the number of entries through the CognitionWall before mice reached the learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). The number of entries to reach this criterion is plotted in a survival plot, with log rank statistics, in order to be able to include mice that did not reach the learning criterion during the task.

The primary outcome measure to investigate general activity differences is the

__total number of entries__through any of the holes during the reversal learning phase. Individual mice that make very few entries, and therefore do not experience the operant schedule to the extent that produces learning, should not be included in the evaluation of reversal learning.

**Parameter information**

Parameter name | Units | Explanation |

Entries to criterion Initial Discrimination Learning | Entries to criterion (count) | The total number of entries through any of the three holes in the CognitionWall is counted, until the mice have reached the learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). This bar graph displays on the y-axis the number of entries to reach the different criterion levels (x-axis) ranging from 70, 73, 77, 80, 83, 87 to 90 percent correct during the initial learning phase. It is advisable to use the 80% criterion as learning criterion, since lower performance criteria can be reached by chance, and higher performance criteria may take very long to be reached (if reached at all) since mice have a tendency to keep on sampling from non-rewarded entrances. Furthermore, since not all mice may reach the learning criterion, it is advisable to use the survival plot with log rank statistics, in order to be able to include mice in the statistical analyses that did not reach the learning criterion. |

Errors to criterion Initial Discrimination Learning | Errors to criterion (count) | The total number of incorrect entries through the middle and right hole (non-rewarded holes) of the CognitionWall are counted, until the mice have reached the learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). This bar graph displays on the y-axis the number of errors until the different criterion levels have been reached (x-axis), ranging from 70, 73, 77, 80, 83, 87 to 90 percent correct during the initial learning phase. |

Entry count per entrance Initial Discrimination Learning (2 days) | Number of entries (count) | Number of entries through the left, middle and right entrances during 2 days of initial learning. |

Entry count per entrance Reversal Learning (2 days) | Entries (count) | Number of entries through the left, middle and right entrances during 2 days of reversal learning. |

Entries to criterion Reversal Learning | Entries to criterion (count) | The total number of entries through any of the three holes in the CognitionWall is counted, until the mice have reached the reversal learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). This bar graph displays on the y-axis the number of entries to reach the different criterion levels (x-axis) ranging from 70, 73, 77, 80, 83, 87 to 90 percent correct during the reversal learning phase. It is advisable to use the 80% criterion as learning criterion, since lower performance criteria can be reached by chance, and higher performance criteria may take very long to be reached (if reached at all) since mice have a tendency to keep on sampling from non-rewarded entrances. Furthermore, since not all mice may reach the learning criterion, it is advisable to use the survival plot with log rank statistics, in order to be able to include mice in the statistical analyses that did not reach the learning criterion. |

Errors to criterion Reversal Learning | Errors to criterion (count) | The total number of incorrect entries through the middle and left hole (non-rewarded holes) of the CognitionWall are counted during the reversal learning phase of the task, until the mice have reached the reversal learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). This bar graph displays on the y-axis the number of errors until the different criterion levels have been reached (x-axis), ranging from 70, 73, 77, 80, 83, 87 to 90 percent correct during the initial learning phase. |

Side bias | Number of entries (count) | Total number of entries through the left, middle and right entrance during the first 30 entries after protocol start. This measure is used to determine if mice have a systematic bias to go through one of the three entrances before learning. Mice tend to visit the metal food grid in the cage frequently (which in most facilities is located next to the right entrance), typically resulting in a side bias towards the right entrance. Therefore, during initial learning, the left entrance in deemed the correct (rewarded) entrance. |

*Pellet drops Initial Discrimination Learning (2 days) | Pellets (count) | The total number of pellets (food rewards) earned during the initial discrimination learning phase. |

Total entries Initial Discrimination Learning | Entries (count) | The total number of entries made during the initial learning phase through any of the holes (left, middle and right) provides a measure of general task activity. Individual mice that make very few entries, and therefore do not experience the operant schedule to the extent that produces learning, should not be included in the evaluation of initial discrimination learning. |

Hours to criterion Initial Discrimination Learning | Time to criterion (hours) | The time, in hours rather than entries, needed to reach the different criterion levels during the initial learning phase. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). This bar graph displays on the y-axis the hours to reach the different criterion levels (x-axis) ranging from 70, 73, 77, 80, 83, 87 to 90 percent correct during the initial discrimination learning phase. It is advisable to use the 80% criterion as learning criterion, since lower performance criteria can be reached by chance, and higher performance criteria may take very long to be reached (if reached at all) since mice have a tendency to keep on sampling from non-rewarded entrances. Furthermore, since general task activity heavily influences this measure, it is advisable to use the "number of entries to reach the criterion" rather than this duration parameter to reach the criterion. |

*Pellet drops Reversal Learning | Pellets (count) | The total number of pellets (food rewards) earned during the reversal learning phase. |

Total entries Reversal Learning | Entries (count) | The total number of entries made during the reversal learning phase through any of the holes (left, middle and right) provides a measure of general task activity. Individual mice that make very few entries, and therefore do not experience the operant schedule to the extent that produces learning, should not be included in the evaluation of reversal learning. |

Hours to criterion Reversal Learning | Time to criterion (hours) | The time, in hours rather than entries, needed to reach the different criterion levels during the reversal learning phase. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). This bar graph displays on the y-axis the hours to reach the different criterion levels (x-axis) ranging from 70, 73, 77, 80, 83, 87 to 90 percent correct during the reversal learning phase. It is advisable to use the 80% criterion as learning criterion, since lower performance criteria can be reached by chance, and higher performance criteria may take very long to be reached (if reached at all) since mice have a tendency to keep on sampling from non-rewarded entrances. Furthermore, since general task activity heavily influences this measure, it is advisable to use the "number of entries to reach the criterion" rather than this duration parameter to reach the criterion. |

Survival plot of entries to 70% criterion Initial Discrimination Learning | Mice that reached criterion (fraction) | The total number of entries through any of the three holes in the CognitionWall is counted, until the mice have reached this learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). Each "step" in the curve represents a mouse that reached the learning criterion at the respective number of entries (x-axis). From this plot, it is evident when mice did not reach the learning criterion (i.e. the fraction of mice reaching the criterion (y-axis) does not reach the value of 1). Even though not all mice may reach the learning criterion (due to poor learning performance), these mice are still included in the log rank statistics, including poor learning mice in the statistical analyses. |

Survival plot of entries to 83% criterion Initial Discrimination Learning | Mice that reached criterion (fraction) | The total number of entries through any of the three holes in the CognitionWall is counted, until the mice have reached this learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). Each "step" in the curve represents a mouse that reached the learning criterion at the respective number of entries (x-axis). From this plot, it is evident when mice did not reach the learning criterion (i.e. the fraction of mice reaching the criterion (y-axis) does not reach the value of 1). Even though not all mice may reach the learning criterion (due to poor learning performance), these mice are still included in the log rank statistics, including poor learning mice in the statistical analyses. |

Survival plot of entries to 73% criterion Initial Discrimination Learning | Mice that reached criterion (fraction) | The total number of entries through any of the three holes in the CognitionWall is counted, until the mice have reached this learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). Each "step" in the curve represents a mouse that reached the learning criterion at the respective number of entries (x-axis). From this plot, it is evident when mice did not reach the learning criterion (i.e. the fraction of mice reaching the criterion (y-axis) does not reach the value of 1). Even though not all mice may reach the learning criterion (due to poor learning performance), these mice are still included in the log rank statistics, including poor learning mice in the statistical analyses. |

Survival plot of entries to 77% criterion Initial Discrimination Learning | Mice that reached criterion (fraction) | |

*Survival plot of entries to 80% criterion Initial Discrimination Learning | Mice that reached criterion (fraction) | |

Survival plot of entries to 87% criterion Initial Discrimination Learning | Mice that reached criterion (fraction) | |

Survival plot of entries to 90% criterion Initial Discrimination Learning | Mice that reached criterion (fraction) | |

Survival plot of entries to 70% criterion Reversal Learning | Mice that reached criterion (fraction) | |

Survival plot of entries to 73% criterion Reversal Learning | Mice that reached criterion (fraction) | |

Survival plot of entries to 77% criterion Reversal Learning | Mice that reached criterion (fraction) | |

*Survival plot of entries to 80% criterion Reversal Learning | Mice that reached criterion (fraction) | |

Survival plot of entries to 83% criterion Reversal Learning | Mice that reached criterion (fraction) | |

Survival plot of entries to 87% criterion Reversal Learning | Mice that reached criterion (fraction) | |

Survival plot of entries to 90% criterion Reversal Learning | Mice that reached criterion (fraction) | |

*Number of entries required to 80% criterion Initial Discrimination Learning | Entries to criterion (count) | The total number of entries through any of the three holes in the CognitionWall is counted, until the mice have reached the learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). Since not all mice may reach the learning criterion, it is advisable to use the survival plot with log rank statistics, in order to be able to include mice in the statistical analyses that did not reach the learning criterion. |

*Number of entries to 80% criterion Reversal Learning | Entries to criterion (count) | The total number of entries through any of the three holes in the CognitionWall is counted, until the mice have reached the learning criterion. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). Since not all mice may reach the learning criterion, it is advisable to use the survival plot with log rank statistics, in order to be able to include mice in the statistical analyses that did not reach the learning criterion. |

*Pellet drops Initial Discrimination Learning (1 day) | Pellets (count) | The number of pellets (food rewards) earned during the first day of initial discrimination learning phase. |

Entry count per entrance Initial Discrimination Learning (1 day) | Number of entries (count) | Number of entries through the left, middle and right entrances during the first day of initial learning. |

*Pellet drops Initial Discrimination Learning (3 days) | Pellets (count) | The total number of pellets (food rewards) earned during the initial discrimination learning phase. |

Entry count per entrance Initial Discrimination Learning (3 days) | Number of entries (count) | Number of entries through the left, middle and right entrances during 3 days of initial learning. |

*Pellet drops Reversal Learning (3 days) | Pellets (count) | The total number of pellets (food rewards) earned during the reversal learning phase. |

Entry count per entrance Reversal Learning (3 days) | Entries (count) | Number of entries through the left, middle and right entrances during 3 days of reversal learning. |

Hours to 80% correct criterion Initial Discrimination Learning | Time to criterion (hours) | The time, in hours rather than entries, needed to reach the different criterion levels during the initial learning phase. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). This bar graph displays on the y-axis the hours to reach the different criterion levels (x-axis) ranging from 70, 73, 77, 80, 83, 87 to 90 percent correct during the initial discrimination learning phase. It is advisable to use the 80% criterion as learning criterion, since lower performance criteria can be reached by chance, and higher performance criteria may take very long to be reached (if reached at all) since mice have a tendency to keep on sampling from non-rewarded entrances. Furthermore, since general task activity heavily influences this measure, it is advisable to use the "number of entries to reach the criterion" rather than this duration parameter to reach the criterion. |

Hours to 80% correct criterion Reversal Learning | Time to criterion (hours) | The time, in hours rather than entries, needed to reach the different criterion levels during the reversal learning phase. This learning criterion is calculated as the percentage of correct entries of the last 30 entries (i.e. moving window of the last 30 entries). This bar graph displays on the y-axis the hours to reach the different criterion levels (x-axis) ranging from 70, 73, 77, 80, 83, 87 to 90 percent correct during the initial discrimination learning phase. It is advisable to use the 80% criterion as learning criterion, since lower performance criteria can be reached by chance, and higher performance criteria may take very long to be reached (if reached at all) since mice have a tendency to keep on sampling from non-rewarded entrances. Furthermore, since general task activity heavily influences this measure, it is advisable to use the "number of entries to reach the criterion" rather than this duration parameter to reach the criterion. |

Initial learning performance after 150 entries | Performance (ratio) | |

Initial learning performance after 200 entries | Performance (ratio) | |

Initial learning performance after 250 entries | Performance (ratio) | |

Initial learning average performance between 100 - 150 entries | Performance (ratio) | |

Initial learning average performance between 150 - 200 entries | Performance (ratio) | |

Initial learning average performance between 200 - 250 entries | Performance (ratio) | |

Reversal learning performance after 150 entries | Performance (ratio) | |

Reversal learning performance after 200 entries | Performance (ratio) | |

Reversal learning performance after 250 entries | Performance (ratio) | |

Reversal learning average performance between 100 - 150 entries | Performance (ratio) | |

Reversal learning average performance between 150 - 200 entries | Performance (ratio) | |

Reversal learning average performance between 200 - 250 entries | Performance (ratio) | |

*Pellet drops Initial Discrimination Learning (4 days) | Pellets (count) | The total number of pellets (food rewards) earned during the initial discrimination learning phase. |

Entry count per entrance Initial Discrimination Learning (4 days) | Number of entries (count) | Number of entries through the left, middle and right entrances during 4 days of initial learning. |

Number of perseverative errors to 70% criterion reversal | Frequency | Number of perseverative errors to 70% correct during reversal (window 30) |

Number of perseverative errors to 73% criterion reversal | Frequency | Number of perseverative errors to 73% correct during reversal (window 30) |

Number of perseverative errors to 77% criterion reversal | Frequency | Number of perseverative errors to 77% correct during reversal (window 30) |

Number of perseverative errors to 80% criterion reversal | Frequency | Number of perseverative errors to 80% correct during reversal (window 30) |

Number of perseverative errors to 83% criterion reversal | Frequency | Number of perseverative errors to 83% correct during reversal (window 30) |

Number of perseverative errors to 87% criterion reversal | Frequency | Number of perseverative errors to 87% correct during reversal (window 30) |

Number of perseverative errors to 90% criterion reversal | Frequency | Number of perseverative errors to 90% correct during reversal (window 30) |

Number of neutral errors to 70% criterion reversal | Frequency | Number of neutral errors to 70% correct during reversal (window 30) |

Number of neutral errors to 73% criterion reversal | Frequency | Number of neutral errors to 73% correct during reversal (window 30) |

Number of neutral errors to 77% criterion reversal | Frequency | Number of neutral errors to 77% correct during reversal (window 30) |

Number of neutral errors to 80% criterion reversal | Frequency | Number of neutral errors to 80% correct during reversal (window 30) |

Number of neutral errors to 83% criterion reversal | Frequency | Number of neutral errors to 83% correct during reversal (window 30) |

Number of neutral errors to 87% criterion reversal | Frequency | Number of neutral errors to 87% correct during reversal (window 30) |

Number of neutral errors to 90% criterion reversal | Frequency | Number of neutral errors to 90% correct during reversal (window 30) |