Memory-guided navigation in amyotrophic lateral sclerosis

Background Previous studies have yielded inconsistent results about hippocampal involvement in non-demented patients with amyotrophic lateral sclerosis (ALS). We hypothesized that testing of memory-guided spatial navigation i.e., a highly hippocampus-dependent behaviour, might reveal behavioural correlates of hippocampal dysfunction in non-demented ALS patients. Methods We conducted a prospective study of spatial cognition in 43 non-demented ALS outpatients (11f, 32 m, mean age 60.0 years, mean disease duration 27.0 months, mean ALSFRS-R score 40.0) and 43 healthy controls (14f, 29 m, mean age 57.0 years). Participants were tested with a virtual memory-guided navigation task derived from animal research (“starmaze”) that has previously been used in studies of hippocampal function. Participants were further tested with neuropsychological tests of visuospatial memory (SPART, 10/36 Spatial Recall Test), fluency (5PT, five-point test) and orientation (PTSOT, Perspective Taking/Spatial Orientation Test). Results Patients successfully learned and navigated the starmaze from memory, both in conditions that forced memory of landmarks (success: patients 50.7%, controls 47.7%, p = 0.786) and memory of path sequences (success: patients 96.5%, controls 94.0%, p = 0.937). Measures of navigational efficacy (latency, path error and navigational uncertainty) did not differ between groups (p ≥ 0.546). Likewise, SPART, 5PT and PTSOT scores did not differ between groups (p ≥ 0.238). Conclusions This study found no behavioural correlate for hippocampal dysfunction in non-demented ALS patients. These findings support the view that the individual cognitive phenotype of ALS may relate to distinct disease subtypes rather than being a variable expression of the same underlying condition. Supplementary Information The online version contains supplementary material available at 10.1007/s00415-023-11753-8.

Supplementary table 1. Block, conditions and trial order. L denotes learning trials, where the goal appeared at the target location. P denotes probe trial, where the goal did not appear and the participant had to press a button at the chosen target location. Video was only presented if the participant did not find the goal during the first three trials. In baseline trials, environmental landmarks were visible and participants started from the same starting point. In egocentric trials, environmental landmarks were invisible and participants started from the same starting point. In allocentric trials, environmental landmarks were visible and participants started from a new starting point.

Block Condition
Trial order Number of trials 1 baseline Supplementary material on data pre-processing Success rate: We determined whether our subjects successfully navigated to the correct location in probe trials, by calculating the Euclidean distance between the target and the participant's final position.
We classified a trial as successful if the Euclidean distance was less than one-third of the length of an external maze alley. We then calculated the percentage of successful trials.
Latency: For successful probe trials, we calculated the trial duration by subtracting the first timestamp from the last timestamp.
Path error: For successful probe trials, we determined the total distance travelled by calculating both the participant's path length and the ideal path length to the target location. This measure reflects the degree of directness of navigation. We normalized the path length by calculating the absolute percent error. ( Search accuracy: For successful probe trials, we determined the average distance to the target across all time stamps of a trial to determine how close to the target participants searched. This measure does not correlate with path length per se, but rather refers to the degree of uncertainty in navigation behaviour, even on trials with normal path length. We normalized the distance to target by calculating the absolute percent error. (Ideal) distance to target =

Ratings of post-navigational memory
The maze reconstruction task and landmark identification task was rated by three independent examiners according to predefined criteria. Criteria for maze reconstruction: 1. evaluation of outer shape (1 point each): symmetric, closed figure, intersections with at least 2 choice points, single alley as connector between intersections, equal alley length, extra point for correct outer shape; 2. evaluation of inner shape (1 point each): singular structure, closed, squared, has 5 edges; 3. number of peripheral alleys (1 point each, minus if more than 5 alleys). The points for all three aspects were added and normalized to a score from zero to one; higher score denotes more accurate maze reconstruction. Criteria for landmark identification: Number of landmarks (1 point each, minus if more than 5 landmarks), accuracy of semantic categories of landmarks (1 point each, minus for category duplicates), 3. level of detail (1 point each): 4 electrical towers, mountain village, second mountain village. The points for all three aspects were added and normalized to a score from zero to one; higher score denotes better landmark memory. To obtain final scores for each subject, the three examiner ratings were averaged.
For quantifying the positioning of environmental landmarks, we used the Gardony Map Drawing Analyzer software [34]. As recommended by the authors, we computed the SQRT(CanOrg) score, by determining pairwise canonical relations (north-south/east-west) between objects in the sketch map and comparing them to the target map. The SQRT(CanOrg) score ranges between zero and one with higher values denoting better overall configural accuracy and completeness.
Supplementary figure 1. Example trajectories in selected probe trials. First row, baseline condition, second row egocentric condition, third row allocentric condition. The first column shows the ideal path from the start to the target, followed by the actual trajectories of the ALS group (blue), followed by the actual trajectories of the control group (yellow).  , 2017) i.e. verbal fluency, executive function or language function (ECAS-), controls with normal ECAS scores (ECAS+), controls with ECAS scores below cut-off in fluency, executive or language function (ECAS-). Data presented as median with interquartile range (25 -75%). Values for latency, path error and search accuracy based on successful probe trials. Statistical analysis with Kruskal-Wallis rank sum test. Abbreviations: ALS = amyotrophic lateral sclerosis; Ctrl = controls. Note comparable memoryguided spatial navigation performance for all investigated parameters across groups.  .406

ALS-ECAS+
Supplementary table 5. Performance in the starmaze task for patients with ALS, PLS, PMA, and controls. Data presented as median with interquartile range (25 -75%). Values for latency, path error and search accuracy based on successful probe trials only. Statistical analysis with Kruskal-Wallis rank sum test. Abbreviations: ALS = amyotrophic lateral sclerosis; PLS = primary lateral sclerosis, PMA = progressive muscular atrophy. Note that all groups show comparable memory-guided spatial navigation performance, as indicated by the lack of group differences in success rate, latency, path error, and search accuracy.