Trial-unique, delayed nonmatching-to-location (TUNL) touchscreen testing for mice: sensitivity to dorsal hippocampal dysfunction

Abstract

Rationale: The hippocampus is implicated in many of the cognitive impairments observed in conditions such as Alzheimer's disease (AD) and schizophrenia (SCZ). Often, mice are the species of choice for models of these diseases and the study of the relationship between brain and behaviour more generally. Thus, automated and efficient hippocampal-sensitive cognitive tests for the mouse are important for developing therapeutic targets for these diseases, and understanding brain-behaviour relationships. One promising option is to adapt the touchscreen-based trial-unique nonmatching-to-location (TUNL) task that has been shown to be sensitive to hippocampal dysfunction in the rat.

Objectives: This study aims to adapt the TUNL task for use in mice and to test for hippocampus-dependency of the task.

Methods: TUNL training protocols were altered such that C57BL/6 mice were able to acquire the task. Following acquisition, dysfunction of the dorsal hippocampus (dHp) was induced using a fibre-sparing excitotoxin, and the effects of manipulation of several task parameters were examined.

Results: Mice could acquire the TUNL task using training optimised for the mouse (experiments 1). TUNL was found to be sensitive to dHp dysfunction in the mouse (experiments 2, 3 and 4). In addition, we observed that performance of dHp dysfunction group was somewhat consistently lower when sample locations were presented in the centre of the screen.

Conclusions: This study opens up the possibility of testing both mouse and rat models on this flexible and hippocampus-sensitive touchscreen task.

Keywords: Delayed nonmatching-to-location; Hippocampus; Mouse; Spatial pattern separation; Spatial working memory; Touchscreen operant chamber.

Fig. 1

The TUNL task. The screen is covered with a black Perspex mask with five response windows. In the choice phase, the correct stimulus (white square) denoted by “+” is in a new location that does not match the sample location denoted by “−”. This trial has a spatial separation level of 3 (S3), i.e. number of response windows between the two stimuli is 3 (see Table 1)

Fig. 2

Rate of acquisition of mouse TUNL training, and sensitivity of performance to separation and delay. a Number of sessions taken to reach the criterion for each spatial separation level in stage 1; b effect of separation in stage 2 under mixed S1 and S0 condition (0 s delay); c effect of delay in stage 2 (mixed S1 and S0). Data are presented as mean ± standard error of the mean (SEM). **p < .005

Fig. 3

Photographs of representative sections and diagram of the extent of dorsal hippocampal lesions. a Photographs of coronal sections corresponding to −1.7 mm from bregma; b light grey represents the smallest lesion and dark grey represents the largest. All numbers correspond to distance in millimetre from bregma

Fig. 4

Effects of hippocampal lesions on TUNL. a Post surgery re-acquisition in stage 2 (S1, 0 s delay); b effects of delay in two mixed-delay conditions in stage 2 (S1); c effects of spatial separation in a mixed S1 and S0 condition in stage 2 (0 s delay); d detailed analysis of c by sample location. Data are presented as mean ± standard error of the mean (SEM). ns denotes not significant, **p < .005 main effect of lesion

Fig. 5

Mean latencies during the post surgery re-acquisition without programmed delay (0 s delay). The “Sample touch to Choice touch” data in the middle indicates there was an approximately 4 s latency between the sample and choice responses under the 0-s programmed delay condition. This was due to travel time between initiation and screen response. Data are presented as mean ± standard error of the mean (SEM). ns denotes not significant

Fig. 6

Effects of hippocampal lesion on TUNL (stage 3), when predictability of the correct location was low. Data are presented as mean ± standard error of the mean (SEM). **p < .005 main effect of lesion

Fig. 7

Effects of hippocampal lesion on TUNL (stage 1), when predictability of the correct location was high. a Initial mixed-delay sessions (0 and 6 s mixed); b after extended training in two mixed-delay conditions (0, 6, 9 s mixed; 0, 9, 18 s mixed); c two mixed spatial separation conditions (S3 and S2 mixed; S3 and S1 mixed). Data are presented as mean ± standard error of the mean (SEM). ns denotes not significant, *p < .05 main effect of lesion