Juvenile and adult sea Lamprey behaviour (twitch and movement)

Sea lamprey are invasive fish species in the Laurentian Great Lakes and parasitically feed on many commercially important fishes. Sound has been used as a deterrent for invasive species, but its potential for manipulating sea lamprey behaviour in natural stream conditions is under-tested. Both the behavioural acoustic threshold and differential responses at different life stages of sea lamprey have also yet to be established. To fill in some of these literature gaps, low-frequency tones of 70 or 90 Hz were used in a laboratory setting to determine the behavioural responses of adult and juvenile sea lamprey. Both stages of sea lamprey exhibited a change in swimming behaviour and a twitch (startle) in response to both frequencies. Adults had a lower threshold to 70 Hz than 90 Hz tone bursts with juveniles showing no difference between frequencies. These thresholds and behaviours can be used in natural settings to control or impact the behaviour of sea lamprey to aid in directing lamprey movement either toward existing traps or away from fish bypass structures.

Methods
Data was collected using experimental tones, of 70 or 90 Hz, played at varying amplitude levels (level 0-15 on the MP3 player, 129 – 169 dB re 1 µPa, -29.95 to -34.75 dB re 1 m/s2). To determine behavioural reactions and thresholds, sea lamprey would need to be consistently still for 1 minute; sound would then be played for 1 minute, and behavioural reactions (i.e. movement/swimming) were recorded during the sound treatment; this was repeated with multiple sessions (1 minute of no movement followed by 1 minute of sound exposure) within one trial. The behavioural reaction of the sea lamprey was also recorded directly after (1-10 seconds) the sound was played for 1 minute (swimming needed to start within 1-10 seconds but would be recorded for 1 minute to look at proportion of the time spent swimming). A stepwise method was used for decibel changes between sessions within a trial, where sea lamprey were started at a low decibel level sound and no sea lamprey movement (i.e. swimming) within the minute of sound exposure, would lead to a higher decibel level for the next session. Sea lamprey movement response (i.e. swimming) would lead to a lower decibel levels or a repeated decibel level (if there was a question of small movement and needed to be re-checked), similar to methods used and modified in other studies (Tavolga and Wodinsky 1965; Lu et al. 1996; Hawkins and Chapman 2020). Sessions within a trial would continue until sea lamprey stopped behaviourally responding (swimming), potentially resulting from habituation or exhaustion or until the highest decibel level was reached. 

The immediate response to sound application was also used to compare the behaviour of adult and juvenile sea lamprey. This “twitch” response, defined by a sudden, short-term tail or body movement that occurred within the first second of sound presentation, and was quantified as present or not present for the first presentation of any decibel level for each individual. The twitch response was compared to the individual sea lamprey’s baseline behaviour during the one-minute period immediately preceding the sound. Analysis of all trials examined the proportion of time spent swimming, prior (no treatment), during (onset of sound) and after (offset of sound) noise treatments.

In total, 54 sea lamprey were used in this experiment, 22 juvenile sea lamprey and 32 adult sea lamprey. 32 trials were performed on adult sea lamprey with 17 trials with 70 Hz and 15 trials with 90 Hz. 22 trials were performed on juvenile sea lamprey, with 10 trials with 70 and 11 trials with 90 Hz tones. Five trials included both frequencies, three trials started with 90 Hz, followed by high amplitude 70Hz tones (level 12-15; 161-169 dB re 1 µPa), and two trials started with 70 Hz followed by high amplitude 90 Hz tones (level 12-15; 160-167 dB re 1 µPa) to compare high amplitude frequencies.

Behavioural threshold levels for movement during sound exposure were determined by quantifying the proportion of time spent swimming; this was compared to the sea lamprey baseline taken from unmoving behaviour needed for a full 1 minute before playing sound. Since the trials started with lower decibel levels which were followed by increases (if no movement was detected), the behavioural threshold was quantified as the first decibel level of either frequency used (70 or 90 Hz) that elicited a “positive response” behaviour during sound. To determine the threshold of movement during sound playing, we examined the time spent swimming during the one minute of sound for a “positive response” if a sea lamprey was moving more than or equal to 50% of the time (30 sec of 60 sec or 0.50 proportionally) then it was characterized as a “positive response”. To determine the behavioural threshold of movement directly after sound had stopped playing, we similarly examined the proportion of time spent swimming after the sound for a “positive response,” defined as a sea lamprey that moved within the first 10 seconds after the offset. A behavioural threshold for movement after sound has ceased is harder to determine and therefore, we used more resilient standards for “positive response”; movement after sound must have been more than or equal to 80% of the one minute period (48 sec of 60 or 0.80 proportionally) and the sea lamprey must not have made any movement during the 1 minute of sound exposure. The two separate thresholds (movement during sound and movement after sound) were then determined by examining all positive responses and finding the first and the lowest sound level leading to, maximum, one threshold per sea lamprey for during sound and after sound. Level 0 was not included as a threshold even if the behavioural response was classified as “positive response” due to no sound change from background noise, this was assumed to be random behaviour (one level 3 (133 re 1 µPa, -30.43 dB re 1 m/s2) was not included because it was first level played of the session and behavioural response was 0.5 and we were unable to determine if it was a behavioural threshold or random behaviour). The sound levels (determined for each responding sea lamprey) were classified as the dependent variable, while frequency (70 and 90 Hz) and life stage (adult and juvenile) were the independent variables, using an ANOVA in SPSS (IBM).

The twitch response of lamprey to sound was also used to determine the behavioural thresholds. The proportion of twitches were calculated for each level (0-15, 129-169 dB re 1 µPa, -29.95 to -34.75 dB re 1 m/s2) for each life stage (adults and juveniles) and frequency (70 and 90 Hz) by the proportion of individual sea lamprey that presented the twitch behaviour to the total number of individuals exposed to the sound decibel within that life stage and frequency. The proportion of twitches was classified as the dependent variable in adults and juveniles, while frequency (70 and 90 Hz), life stage, and differences between the different sound levels (0-15, 129-169 dB re 1 µPa, -29.95 to -34.75 dB re 1 m/s2) were classified as independent variables, which were evaluated using a Kruskal-Wallis test as this data did not follow a normal distribution. To test the difference between levels, arbitrary numbers (0-15, based on volume levels of the Mp3 player) were used in statistics, which could then be translated to decibel level using Table 1. Sound levels that were presented to all sea lamprey less than three times in total during all trials, were not considered within the life stage and frequency tested (adult and juvenile, 70 or 90 Hz) since all provided insufficient data for statistical analysis. Further insight into the data was needed, so a separate test specifically comparing each level proportion (1-15; 129-169 dB re 1 µPa, -29.95 to -34.75 dB re 1 m/s2) with level 0 (no sound/129 dB re 1 µPa, -29.95 dB re 1 m/s2), was used to statistically determine which levels were significantly different from level 0 or no sound. Comparisons of levels to level 0 (no movement) were evaluated with simple contrasts relative to level 0, with a statistical difference of p<0.05 showing a difference from zero for each level tested against.