The impact of anthropogenic noise on individual identification via female song in Black-capped chickadees (Poecile atricapillus)

Subjects

In total, twenty-two black-capped chickadees (nine males and 13 females) were tested between May and December 2019, and 16 black-capped chickadees (seven males and nine females) completed the experiment. One male and one female failed to learn Pretraining, and one female failed to learn Non-differential training (see descriptions below for training information); as a result, all three were removed from the experiment. In addition, one male and two females died of natural causes during the course of the study (see Ethics Declaration). For all birds, sex was determined by deoxyribonucleic acid analysis of blood samples³⁷. All birds were captured in Edmonton (North Saskatchewan River Valley, 53.53°N, 113.53°W; Mill Creek Ravine, 53.52°N, 113.47°W), Alberta, Canada in January 2018 and January 2019 and were at least one year of age at capture, verified by examining outer tail rectrices³⁸.

Prior to the current experiment, all chickadees were individually housed in Jupiter Parakeet cages (30 × 40 × 40 cm; Rolf C. Hagen, Inc., Montreal, QB, Canada) in a single colony room. Therefore, birds did not have physical contact with each other, but did have visual and auditory contact. Birds had ad libitum access to food (Mazuri Small Bird Maintenance Diet; Mazuri, St. Louis, MO, USA), water with vitamins supplemented on alternating days (Monday, Wednesday, Friday; Prime Vitamin Supplement; Hagen, Inc.), a cup containing grit, and a cuttlebone. Additional nutritional supplements included three to five sunflower seeds daily, one superworm (Zophobas morio) three times a week, and a mixture of hard-boiled eggs and greens (spinach or parsley) twice a week. The colony rooms were maintained at approximately 20 °C and on a light:dark cycle that followed the natural light cycle for Edmonton, Alberta, Canada.

One bird had previous experience with one operant experiment involving chick-a-dee calls but showed no difference in responding in comparison to the naïve birds. The remaining 15 birds had no previous experimental experience with black-capped chickadee-produced fee-bee songs or any experimental paradigm.

Recordings of acoustic stimuli

The following acoustic stimuli were used in our previous published operant study which indicated that male and female chickadees can identify individual females via their song³⁶. Stimuli included the songs of six female black-capped chickadees. All females were captured in Edmonton (North Saskatchewan River Valley, 53.53°N, 113.53°W; Mill Creek Ravine, 53.52°N, 113.47°W), Alberta, Canada in January 2010, 2011, 2012, and 2014, and all females were at least one year of age at capture, verified by examining outer tail rectrices³⁸. Four females were recorded in Spring 2012 and two females were recorded in Fall 2014. Each recording session lasted approximately 1 h and all recordings took place after colony lights turned on at 08:00, specifically at 8:15. All females were recorded in silence, individually, within their respective colony room cages. Colony room cages were placed in sound-attenuating chambers for recording (1.7 m × 0.84 m × 0.58 m; Industrial Acoustics Company, Bronx, NY). An AKG C 1000S (AKG Acoustics, Vienna, Austria) microphone (positioned 0.1 m above and slightly behind the cage) was connected to a Marantz PMD670 (Marantz America, Mahwah, NJ) digital recorder (16 bit, 44,100 Hz sampling rate) and was used for all recordings. Audio recordings were analyzed and cut into individual files (songs) using SIGNAL 5.03.11 software (Engineering Design, Berkley, CA, USA).

Acoustic stimuli

For the current study, a total of 150 vocalizations were used as stimuli, these vocalizations were comprised of 25 fee-bee songs produced by each of six recorded female chickadees. We ensured that all 150 were of high quality, meaning no audible interference, and all stimuli were bandpass filtered (lower bandpass 500 Hz, upper bandpass 14,000 Hz) using GoldWave version 6.31 (GoldWave, Inc., St. John’s, NL, Canada) in order to reduce any background noise outside of the song stimuli spectrum. For each song stimulus, 5 ms of silence was added to the leading and trailing portion of the vocalization and each stimulus was tapered to remove transients, in addition amplitude was equalized peak to peak using SIGNAL 5.03.11 software. When triggered, stimuli were presented at approximately 75 dB peak SPL as measured by a calibrated Brüel & Kjær Type 2239 (Brüel & Kjær Sound & Vibration Measurement A/S, Nærum, Denmark) sound pressure meter (A-weighting, slow response), a level that corresponds with the natural chickadee vocalizations amplitudes^39,40,41. All dB measurements were made at the level of the request perch where birds trigger stimuli and where birds are required to remain for the length of the stimuli and all dB measurements refer to SPL.

Noise stimuli

Anthropogenic noise stimuli were originally created and used by Potvin and MacDougall-Shackleton⁴² and by Potvin, Curcio, Swaddle, and MacDougall-Shackleton⁴³. The stimuli were recorded from an urban area in Melbourne, Victoria, Australia and other anthropogenic noise stimuli of various trains, cars, motorcycles, and lawnmowers downloaded from Soundbible.com were used. Within Victoria⁴⁴ and Alberta^45,46, urban traffic noise averages 60–80 dB SPL. The files used varied in length, with those recorded in Melbourne all being 10 min in length and those downloaded from Soundbible.com varying between 1–10 minutes^42,43. In total 10 tracks were used with 30 total minutes of noise stimuli. Three anthropogenic noise conditions were used in the study, including Silence (no noise), Low noise (anthropogenic noise stimuli played at ~ 40 dB peak SPL), and High noise (anthropogenic noise stimuli played at ~ 75 dB peak SPL) replicating the variation of traffic noise experienced in urban areas^42,43. For the Low and High noise conditions the 10 tracks repeated on a randomized loop during data collection (natural light of light/dark cycle) with, thus noise exemplars overlapped songs by chance, to further emulate urban areas. Noise stimuli had natural variations and modulations in frequency and amplitude over the course of the sound files. All dB measurements for noise stimuli included in this study refer to SPL. See Fig. 1 for female song and traffic noise stimuli spectrograms and power spectra.

Apparatus

For the duration of the experiment, birds were housed individually in modified colony room cages (30 × 40 × 40 cm; described above) which were placed inside a ventilated, sound-attenuating operant chamber. See Fig. 2 for illustration of operant conditioning chamber. All chambers were lit with a full spectrum LED bulb (3 W, 250 lm E26, Not-Dim, 5000 K; Lohas LED, Chicago, IL, USA), and maintained the natural light:dark cycle for Edmonton, Alberta. Each cage within each operant chamber contained two perches and an additional perch fitted with an infrared sensor (i.e., the request perch). See Fig. 2C. Each cage also contained a water bottle, grit cup, and cuttlebone See Fig. 2G-2H. Birds had ad libitum access to water (with vitamins supplemented on alternating days; Monday, Wednesday, Friday), grit, and cuttlebone and were provided two superworms daily (a morning and afternoon worm). An opening (11 × 16 cm) located on the left side of the cage allowed the birds to access a motorized feeder, with a red LED light, and equipped with an infrared sensor⁴⁷. See Fig. 2B,D–F. The purpose of the sensor was so that food was only available as a reward for correct responses to auditory stimuli during the operant discrimination task. We should note that performance of the discrimination task is required for access to food and thus maintains motivation. For operation and data collection, a personal computer connected to a single-board computer⁴⁸ scheduled trials and recorded responses to stimuli. Stimuli were played from a personal computer hard drive through a Cambridge Integrated Amplifier (model A300 or Azur 640A; Cambridge Audio, London, England). Data is downloaded once a day in order to reduce stress on subjects as all equipment must be tested following download, requiring contact with subjects. Stimuli played in the chamber through a Fostex full-range speaker (model FE108 Σ or FE108E Σ; Fostex Corp., Japan; frequency response range 80–18,000 Hz) located beside the feeder. See Sturdy and Weisman⁴⁹ for a detailed description of the apparatus. See Fig. 2 for an illustration of the operant conditioning chamber set-up.

Procedure

Operant conditioning

Our current operant conditioning go/no-go set-up is used to understand how birds perceive auditory stimuli. By training the birds to respond to particular stimuli and withhold responding to other stimuli we can compare responses to both types of stimuli. The go/no-go paradigm requires the birds to learn which stimuli require correct responses (go), providing reinforcement (food), and which stimuli require birds to withhold responding (no-go), resulting in the avoidance of punishment (lights out).

The current study follows nine stages, after learning to use the operant conditioning set-up, birds then go through Non-differential training (stage 1) where they will be exposed to all stimuli that will be used in the experiment and to ensure that the birds respond to the stimuli equivalently. Then birds complete Discrimination training (stage 2) where birds on two categories of sounds. One category is rewarded, the other category is punished. Then the Discrimination-85 (stage 3) phase prepares birds for future trials where there is no reward nor punishment. After this point, birds will follow three series (Silence; Low; High) of Discrimination-85 with noise (stage 4, 6, 8) and a corresponding Probe with noise (stage 5, 7, 9), meaning that that each subject will repeated the two discrimination tasks three times with different noise conditions, with the order of noise conditions randomized among individuals. The detailed procedures for each stage are described in the following.

Non-differential training

The purpose of Non-differential training is to engender a high level of responding on all trials, across all stimuli. Once a bird learned to use the request perch fitted with a sensor as well as learned to use the feeder to obtain food then Pretraining began. During Pretraining, birds were trained to respond to a 1 s tone (1,000 Hz) in order to receive access to food. Pretraining occurred over an approximately 15-day period in order to allow acclimatization to the chamber, feeder, and speaker. Following Pretraining was Non-differential training. During Non-differential training, birds received food for responding to all fee-bee song stimuli. All trials began when a bird landed on the request perch and remained on the perch for between 900–1100 ms, at which point a randomly-selected song stimulus played. Songs were presented in random order from trial to trial until all 150 stimuli had been triggered and played without replacement; once all 150 stimuli were played, a new random sequence initiated. In the event that the bird left the request perch during a stimulus presentation, the trial was deemed interrupted, and resulted in a 30 s lights out of the operant chamber. If the bird entered the feeder within 1 s after the stimulus (any stimulus) was played, it was given 1 s access to food, followed by a 30 s intertrial interval. If a bird remained on the request perch during the stimulus presentation and the 1 s following the completion of the stimulus, then the bird received a 60 s intertrial interval with the lights on. Birds continued on Non-differential training until they completed six 450-trial blocks at ≥ 60% responding on average to all stimuli, at least four 450-trial blocks at ≤ 3% difference in responding to future rewarded versus future unrewarded Discrimination stimuli, at least four 450-trial blocks at ≤ 3% difference in responding to future rewarded versus unrewarded Discrimination stimuli. Then following a day of free feed (during which birds had ad libitum access to a food cup) birds completed a second round of Non-differential training in which they completed at least one 450-trial block that met each of the above requirements. A 450-trial block consisted of the bird experiencing each of the 150 stimuli three times. For the current study the average time to complete Non-differential training ranged from 10 to 41 days (M = 21.43, SD = 9).

Discrimination training

Discrimination training procedures included only 114 out of the 150 training stimuli that were previously presented in non-differential training, and responses to these stimuli were now differentially reinforced. Specifically, correct responses to half of the stimuli (“rewarded stimuli”, S+) were positively reinforced with 1 s access to food, and incorrect responses to the other half (“unrewarded stimuli”, S−) were instead punished with a 30-s intertrial interval of lights off within the operant chamber. In regard to criterion, Discrimination training continued until a bird completed six 342-trial blocks with a discrimination ratio between their respective S+ and S− of greater than 0.80 with the last two blocks being consecutive. For discrimination ratio calculations see Response Measures below.

The current subjects were randomly assigned to either a True category discrimination group (n = 10) or Pseudo category discrimination group (n = 6). Furthermore, chickadees in the True category discrimination group were divided into two subgroups: (a) True 1 (n = 5; three females and two males) discriminated between 57 rewarded fee-bee songs produced by three individual female chickadees (S+) and 57 unrewarded fee-bee songs produced by another three individual female chickadees (S−); and (b) True 2 (n = 5; two females and three males) discriminated between the same songs with opposite rewards, properly, the 57 rewarded (S+) fee-bee songs were the S− from True 1 and the 57 unrewarded (S−) fee-bee songs were the S+ from True 1. For birds in the True category discrimination the average number of blocks completed per day for Discrimination training ranged from 2.4–4.4 blocks (3.3 ± 0.7 blocks).

In similitude, the Pseudo category discrimination group was divided into two subgroups: (a) Pseudo 1 (n = 3; two females and one male) discriminated between 57 randomly-selected rewarded (S+) fee-bee songs and 57 randomly-selected unrewarded (S−) fee-bee songs; and (b) the second subgroup Pseudo 2 (n = 3; one female and two males) discriminated between the same songs with opposite rewards, meaning, the 57 rewarded (S+) fee-bee songs were the S− from Pseudo 1 and the 57 unrewarded (S−) fee-bee songs were the S+ from Pseudo 1 (S+) fee-bee songs and 57 randomly-selected unrewarded (S−) fee-bee songs. To explicate, the purpose of the two Pseudo groups was to include a control in which subjects are required to memorize each vocalization independent of the producer rather than be trained to categorize songs according to individual chickadees as the True groups have been. All birds remained in their respective groups (True 1 and 2; Pseudo 1 and 2) for the duration of the study. For birds in the Pseudo category discrimination the average number of blocks completed per day for Discrimination training ranged from 3.3–6.1 blocks (4.34 ± 1.2 blocks).

Discrimination-85 phase

Discrimination-85 was identical to the above Discrimination training except that rewarded songs were reinforced with a reduced probability, P = 0.85. Therefore, for 15% of trials when a rewarded stimulus was played and a bird correctly responded, no access to food was triggered. Instead, a 30 s lights on intertrial interval occurred. The change in reinforcement occurs in order to prepare birds for Probe trials in which novel song stimuli were neither rewarded with access to food nor unrewarded with a lights out, instead nothing occurs. Discrimination-85 continued until birds completed two consecutive 342-trial blocks with a discrimination ratio of at least 0.80.

Discrimination-85 phase with noise

All subjects followed three series (Silence; Low; High) of Discrimination-85 with noise and a corresponding Probe with noise and the order of noise stimuli was randomly-selected for each bird. Discrimination-85 with noise was identical to the Discrimination-85 phase except one of the three noise stimuli conditions (Silence; Low noise, 40 dB SPL; High noise, 75 dB SPL) was played over the song stimuli. The noise stimuli condition was randomly-selected for each bird. Each bird went through three series of Discrimination-85 with noise (Silence; Low; High) until reaching criteria: two consecutive 342-trial blocks with a discrimination ratio of at least 0.80. Here, we were interested in how the addition of noise would impact discrimination between rewarded and unrewarded female song stimuli.

Probe phase with noise

Following each Discrimination-85 phase with noise was a corresponding Probe phase with noise. During Probe the reinforcement contingencies from Discrimination-85 were maintained. In addition to the 114 stimuli from Discrimination training, this stage included 12 novel fee-bee songs (i.e., Probe stimuli), two from each of the six individual females. For True groups, six of these novel songs were categorized as P + and the other six as P-, based on whether they were produced by the same birds as the S+ or the S− training stimuli. For Pseudo groups, the novel songs were not assigned to categories. For both groups, the 12 novel stimuli were neither rewarded (no food access) nor unrewarded (no lights out). The birds completed six 126-trial blocks in which the 114 familiar discrimination stimuli repeated once per block and the 12 probe sequences played once per block. In addition, one of the three noise stimuli conditions (Silence; Low noise, 40 dB SPL; High noise, 75 dB SPL) was played over the song stimuli, and each bird went through three series of Probe with noise (Silence; Low; High) which corresponded with the birds previous Discrimination-85 phase with noise condition. Thus, all birds completed all three Discrimination-85 phases with noise conditions followed by the corresponding Probe with noise conditions, and the order of noise stimuli condition was randomly-selected for each bird. In Probe phases we are interested if subjects can categorize novel stimuli to previously rewarded or unrewarded female birds.

Response measures

For each 342-block trial during training (Discrimination-85 with noise; Probe with noise), proportion response was calculated (R + /(N-I)): R + represents the number of trials in which the bird went to the feeder, N represents the total number of trials, and I represents the number of interrupted trials in which the bird left the perch before the entire stimulus played. For Discrimination training and the Discrimination-85 phase, a discrimination ratio was calculated by dividing the mean proportion response to all S+ stimuli by the mean proportion response to S+ stimuli plus the mean proportion response to S− stimuli. A discrimination ratio = 0.50 specifies equal response to rewarded (S+) and unrewarded (S−) stimuli, a discrimination ratio = 1.00 specifies a perfect discrimination between S+ and S− stimuli. We also collected data regarding the number of blocks and days per stage (Discrimination training; Discrimination-85 training with noise) in order to examine the latency of discrimination learning.

Statistical analyses

All statistical analyses were conducted using SPSS (Version 20, Chicago, SPSS Inc.). In order to compare the number of trials needed to reach criterion and the discrimination ratios between True and Pseudo groups for Discrimination Training we conducted an analysis of variance (ANOVA). For Discrimination-85 with noise (Silence, Low noise, High noise), an ANOVA was conducted to compare the number of trials needed to reach criterion and the discrimination ratios between True and Pseudo groups. We also conducted post-hoc tests in order to reveal any sex differences between groups.

And for Discrimination-85 with noise and Probes with noise repeated measures ANOVA was conducted to compare proportion response to training stimuli and probe stimuli between True groups and Pseudo groups. Lastly, we conducted post-hoc tests in order to reveal any differences in the number of trials to reach criterion during Discrimination training and to Discrimination-85 with noise.

Ethics declaration

Throughout the experiment, birds remained in the testing apparatus to minimize the transport and handling of each bird. One male and two female subjects died from natural causes during operant training. Following the experiment, healthy birds were returned to the colony room for use in future experiments.

All procedures were conducted in accordance with the Canadian Council on Animal Care (CCAC) Guidelines and Policies with approval from the Animal Care and Use Committee for Biosciences for the University of Alberta (AUP 1937), which is consistent with the Animal Care Committee Guidelines for the Use of Animals in Research. Birds were captured and research was conducted under an Environment Canada Canadian Wildlife Service Scientific permit (#13-AB-SC004), Alberta Fish and Wildlife Capture and Research permits (#56,066 and #56,065), and the City of Edmonton Parks permit. All methods are reported in accordance with ARRIVE guidelines.

Source: Ecology - nature.com