PLOS ONE RESEARCH ARTICLE Effects of 3 Tesla magnetic resonance imaging exposure on the behavior and orientation of homing pigeons Columba livia domestica Daniel Garcı́a Párraga1,2, Peter L. Tyack ID3*, Vicente Marco-Cabedo1, José Luis Crespo- Picazo1, Xavier Manteca4, Luis Martı́-Bonmatı́ ID5 1 Research Department, Fundación Oceanogràfic de la Comunidad Valenciana, Valencia, Spain, 2 Biology Department, Avanqua-Oceanográfic SL, Valencia, Spain, 3 Sea Mammal Research Unit, Scottish Oceans Institute, School of Biology, University of St Andrews, East Sands, St Andrews, United Kingdom, 4 School of a1111111111 Veterinary Science, Universitat Autònoma de Barcelona, Barcelona, Spain, 5 Medical Imaging Department a1111111111 and Biomedical Imaging Research Group at La Fe University and Polytechnic Hospital and Health Research a1111111111 Institute, Valencia, Spain a1111111111 a1111111111 * plt@st-andrews.ac.uk Abstract OPEN ACCESS Homing pigeons (Columba livia domestica) were used to test whether clinical magnetic res- Citation: Párraga DG, Tyack PL, Marco-Cabedo V, onance (MR) imaging disrupts orientation of animals that sense the earth’s magnetic field. Crespo-Picazo JL, Manteca X, Martı́-Bonmatı́ L Thirty young pigeons were randomly separated into three groups (n = 10/group). Two (2020) Effects of 3 Tesla magnetic resonance groups were anaesthetized and exposed to either a constant (no sequence) or a varying imaging exposure on the behavior and orientation of homing pigeons Columba livia domestica. PLoS (gradient echo and echo planar sequences) magnetic field within a 3 Tesla MR unit for 15 ONE 15(12): e0241280. https://doi.org/10.1371/ minutes. The control group was not exposed to the MR field but shared all other aspects of journal.pone.0241280 the procedure. One day later, animals were released from a site they had never visited, 15 Editor: Wolfgang Rudolf Bauer, km from the home loft. Three weeks after the procedure, animals were released from a dif- Universitatsklinikum Wurzburg, GERMANY ferent unfamiliar site 30 km from the loft. Measured variables included the time to disappear Received: December 16, 2019 from sight (seconds), vanishing bearing (angle), and the time interval from release to enter- Accepted: October 12, 2020 ing the home loft (hours). On first release, the group exposed to varying field gradients dur- ing image acquisition using 2 different standard sequences showed more variability in the Published: December 18, 2020 vanishing bearing compared to the other groups (p = 0.0003 compared to control group), Copyright: © 2020 Párraga et al. This is an open suggesting interference with orientation. Other measures did not show significant differ- access article distributed under the terms of the Creative Commons Attribution License, which ences between groups. On second release, there were no significant differences between permits unrestricted use, distribution, and groups. Our results on homing pigeons show that regular clinical MR imaging exposure reproduction in any medium, provided the original may temporarily affect the orientation of species that have magnetoreception capabilities. author and source are credited. If exposure to MR imaging disrupted processes that are not specific to magnetoreception, Data Availability Statement: All relevant data are then it may affect other species and other capabilities as well. within the paper and its Supporting information files. Funding: PLT was funded by the US Office of Naval Research (https://www.onr.navy.mil) Grant Introduction Number N00014-18-1-2062 and N00014-20-1- Many species, including bacteria, insects, molluscs, crustaceans, fish, turtles, birds, and mam- 2709. The employees of the Fundación Oceanogràfic de la Comunidad Valenciana were mals, use the earth’s geomagnetic field for navigation [1, 2]. Experimental exposure to mag- funded by the Fundación. The funders had no role netic fields has been shown to alter the orientation of a diverse array of animal species: mole- in study design, data collection and analysis, rats [3], sharks [4], honeybees [5], sea turtles [6], cockroaches [7], birds [8]. Clinical magnetic PLOS ONE | https://doi.org/10.1371/journal.pone.0241280 December 18, 2020 1/9 PLOS ONE 3T MR imaging affects orientation of homing pigeons decision to publish, or preparation of the resonance (MR) imaging uses strong magnetic fields (~3 Tesla or 3T) about 60,000 times the manuscript. average strength of the earth’s magnetic field (~50 μT), but the short and long-term effects of Competing interests: The authors have declared MR imaging have not been demonstrated for any animal species. Clinical MR is increasingly that no competing interests exist. used for wildlife medicine in zoological institutions and rehabilitation centers, which raises concern about the effect of MR on wildlife [9]. MR exposure is considered a physiologically safe procedure in human medicine without deleterious effects at standard field strengths up to at least 3T [10–13]. However, laboratory mice exposed to a 5T constant magnetic field for 24 and 48 h ate and drank less, leading to a reduced body weight [14]. Some authors have pointed out that MR-related forces on magnetite particles located in tissues could cause injury [15]. Loggerhead sea turtles (Caretta caretta) exposed to MR imaging for diagnostic purposes react to the presence of varying magnetic field gradients despite being in a deep anesthetic state [16, 17, and personal observations of the authors]. During MR image acquisitions, sea turtles responded by increasing their heart rate and moving their body parts, such as the head and flippers, moving enough to impede the acquisition of quality images. In this study, we aimed to evaluate the extent to which a short exposure to powerful mag- netic fields, such as the ones used in clinical MR, can induce after-effects on normal orienta- tion behavior. The homing pigeon (Columba livia domestica) was selected as a model species as they use the geomagnetic field for orientation with a highly predictable homing behavior [18, 19]. Materials and methods Study birds The experiment was approved by the Ethical Committee of the Fundación Hospital La Fe (reg- istration number, OEBA-1-2016). Birds were provided by a local official breeder (affiliated with the “Real Federación Colombófila Española” society). Thirty juvenile homing pigeons were included in the study, between 4 and 5 months of age, with no gender determination. All individuals were kept under same living conditions, were ringed for identification, and were confirmed to be clinically healthy by a specialized veterinar- ian. During the first months of life the birds were allowed to fly outside of the loft for several hours each day to develop their flight muscles, learn to form flocks, and learn landmarks near the loft. The animals had not participated in any previous release or experiment. We think that the daily exercise flights from the loft were unlikely to have given the young subjects close-up experience with landmarks visible from the release sites, which were 15 and 30 km away. The animals were returned to their owner once the study was finished. MR methodology All pigeons underwent a 2-hour fasting period inside two transport crates (certified by the Homing Pigeon Federation; 100x75x30 cm, each with a capacity of holding 20 pigeons but only used to hold 15 animals), after which they were taken to the MR facility (84.2 km from the loft). The pigeons were then weighed (average weight 350 g) and anesthetized using a mixture of intramuscular ketamine (30 mg/kg) and diazepam (1mg/kg). The 30 anesthetized pigeons were randomly divided into three groups of 10 pigeons each. No relevant differences were observed in age or weight between the three groups. Warm water bottles were used to prevent hypothermia of anesthetized birds in the cool scanning area. The control group was kept in the preparation room and was not exposed to the MR field. The constant field group was exposed for 15 minutes to the main MR magnetic field of 3 Tesla, all at the same time inside the scanner, without running any MR sequence. The clinical MR (Philips DS Achieva, PLOS ONE | https://doi.org/10.1371/journal.pone.0241280 December 18, 2020 2/9 PLOS ONE 3T MR imaging affects orientation of homing pigeons Philips Healthcare, Andover, USA) was operated at 3T with an 80 mT/m MR gradient and a slew rate of 200 T/m/s. Pigeons from the varying field group were placed simultaneously inside the MR gantry for 15 minutes, while 2 sequences were performed: a 3D Gradient Echo (TR = 9.2 ms; TE = 1.2 ms; 150 slices; gradients = 4.5 mT/m; scan time = 3:28 min) and a 2D Echo Planar (TR = 13556 ms; TE = 40 ms; 76 slices and 40 dynamics; gradients = 0.5 mT/m; scan time = 9:15 min). All pigeons were allowed to recover from the anesthesia in an adequately heated room. The recovery time of each individual was recorded, and no problems were observed during recov- ery. One hour after recovery, the animals were placed back in the transport crate with free access to food and water. They stayed at the animal research facilities until the next day for their first release. Pigeon releases Two controlled releases, separated by 22 days in time and 14.2 km in distance, were conducted to assess speed and accuracy of orientation behavior 1 day and 23 days after the MR exposure. Both releases were done on clear days with no overcast and at dawn to avoid exposing animals to excessive heat. The first release site (40˚3’46.97”N; 0˚7’24.30”W), was located 15.4 km away from the loft (39˚59’0.31”N; 0˚16’17.59”W) and the bearing to home from release site 1 was 235 degrees. The second release site (40˚10’5.43”N; 0˚1’19.04”W) was 29.6 km away from the loft, with a home bearing of 226 degrees. The distances were selected to be far enough that the birds could not find the loft by chance and not so far that the birds would get tired due to their lack of experience with long flights and to minimize the risk of predatory attacks. The environ- mental conditions during the first and second releases were as follows: minimum/maximum temperature 20.3/31.6˚C vs. 21.8/36.1˚C, and relative humidity 61% vs. 47%, respectively. Each bird was released individually; no pigeon was released until the previous one had been out of sight for at least 5 minutes to prevent them from following each other. Pigeons from the 3 groups were alternated in the release sequence. Each pigeon was observed with binoculars (Nikon Monarch 7, 8x42) out to a maximum visible distance of approximately 2 km. Each pigeon was observed circling around the release site and then adopting a straight route of flight until it was no longer visible, at which point the compass angle of its vanishing bearing was noted. The exact time of release and time when each subject vanished from sight was recorded. As pigeons were individually identified with a microchipped ring, a reader at the entrance of the loft (TIPES1 MC2100) was used to register the exact time each pigeon entered the home loft. The reader was removed 72 h after release. Finally, 3 response parameters were recorded from each subject: the duration from release to the time the subject vanished from sight flying in a straight line; the compass angle of its vanishing bearing; and the amount of time it took from release to entering the loft. Circular statistics were performed using the circular package of R-Studio [20], except for two statistics. The calculation of the 95% confidence interval of the mean vectors for vanishing bearing was made using the circ_mean function of the circular statistics toolbox for MATLAB [21]. The logic of the statistical test comparing the spread of angular values (circular analogue of homo- scedasticity) for each pair of conditions was taken from [22], where spread is calculated by deviation angles that compare each measurement to the mean vector for that condition. The statistical test used for this was the Kruskal-Wallis multiple comparison (using R program kruskal.test; p-values for release 1 were adjusted with the Benjamini-Hochberg method [23] derived from [24] using R program dunnTest). The probability that each of the 3 groups had a mean vanishing bearing significantly different from the home direction was calculated using SpecMeanTestBoot function test from [20, section 5.3.3]. The probability that the 3 groups PLOS ONE | https://doi.org/10.1371/journal.pone.0241280 December 18, 2020 3/9 PLOS ONE 3T MR imaging affects orientation of homing pigeons share a common mean direction was calculated using YGTestBoot from section 7.2.2 of [20]. All of the non-circular data showed significant deviations from a normal distribution judging by Q-Q plots and the Shapiro-Wilk normality test (using R program shapiro.test). Therefore, non-circular data were log-transformed and evaluated again for deviation from normality. For parameters where the log-transformed non-circular data still deviated significantly from a nor- mal distribution (Release 1 Vanishing Time; Release 2 Vanishing Time, and Release 2 Arrival Time), differences between the 3 groups were tested using the Kruskal-Wallis test using R pro- gram kruskal.test; otherwise (Release 1 Arrival Time) differences were tested by an ANOVA using R program aov.test. Results After the MR exposure, all the pigeons recovered well from anesthesia over 30–45 minutes (with no statistical differences in recovery time between groups), and no unusual behavior was observed. None of the birds died during the experimental procedures at the MR site. Table 1 shows the mean values for the variables measured from the 3 groups at both releases. First release No differences (p = 0. 93) were observed between the 3 groups in the time each pigeon took to vanish from sight as it flew to the loft. None of the groups had a mean vanishing bearing signif- icantly different from the home direction (Fig 1; control group, p = 0.29; constant field group, p = 0.61), although the varying field group had a close-to-significant difference from the home direction (p = 0.07). None of the groups had a mean vanishing bearing significantly different Table 1. Summary of homing data for the different groups. Release from first site (15.4 km away from the loft with a bearing to home of 235 degrees). Group Vanishing Bearings Vanishing Time interval from release to entering the # Birds Returned Time loft Mean Vector Angle ±95% Confidence Mean Vector Seconds N Average Entering Time (±SD) N >24 Did not Interval (deg) Length (Hours:Minutes) hours return Control 238±6 0.99 147±69 10 26:09±28:43 10 4 0 Constant 238±11 0.96 145±80 10 14:22±21:27 10 1 0 field Varying 261±26 0.85 138±41 10 13:57±13:42 9 3 1 field p-value 0.22 0.93 0.58 Release from second site (29.6 km away from the loft with a bearing to home of 226 degrees). Group Vanishing Bearings Vanishing Time interval from release to entering the # Birds Returned Time loft Mean Vector Angle ±95% Confidence Mean Vector Seconds N Average Entering Time (±SD) N >24 Did not Interval (deg) Length (Hours:Minutes) hours return Control 194±18 0.91 151±87 10 7:27±5:05 10 0 0 Constant 206±19 0.90 174 10 8:22±10:01 10 1 0 field ±154 Varying 189±12 0.96 111±22 9 11:22±10:33 8 1 1 field p-value 0.25 0.40 0.60 The mean vector angle ± the 95% confidence interval and mean vector length of the vanishing bearings are given in columns 2–4 respectively. The probability that the 3 groups share a common mean direction is given in column 2 of the p-value row. The average (±SD) is given for non-circular data in columns 4 and 6, along with probability that the means differ across the 3 groups. Column 8 lists the number of birds that took >24 hours to return to their home loft. https://doi.org/10.1371/journal.pone.0241280.t001 PLOS ONE | https://doi.org/10.1371/journal.pone.0241280 December 18, 2020 4/9 PLOS ONE 3T MR imaging affects orientation of homing pigeons Fig 1. Vanishing bearings of pigeons from the first release along with the mean vector for each group. Black = control; Orange = exposed to constant magnetic field in MRI; Red = exposed to varying magnetic field from scanning sequences. Home direction is marked with the light blue line and “H”. https://doi.org/10.1371/journal.pone.0241280.g001 from each other (Table 1 for the comparison of all 3 groups; Table 2 for pairwise comparisons of groups), although there was a trend for the varying field group to be more different than the others on the first release. The deviation angles from the vanishing bearings to the mean bearing for each group showed a significantly higher scatter for pigeons in the varying field condition compared to the constant field (p = 0.01) and the control group (p = 0.0003) (Fig 1; Table 3). The difference Table 2. Results of testing difference in mean vanishing bearings after first and second releases between pairs of experimental conditions, using the YgTestBoot function from [20]. Condition 1 Condition 2 Release Statistic p-value Control Constant field 1 0.008 0.93 2 1.27 0.29 Control Varying field 1 3.79 0.11 2 0.284 0.62 Constant field Varying field 1 3.52 0.11 2 2.94 0.12 The p-values from this test indicate the probability that the two groups share a common mean direction. https://doi.org/10.1371/journal.pone.0241280.t002 PLOS ONE | https://doi.org/10.1371/journal.pone.0241280 December 18, 2020 5/9 PLOS ONE 3T MR imaging affects orientation of homing pigeons Table 3. Statistical test comparing the spread of vanishing bearings (circular analog of homoscedasticity) for each pair of conditions in Release 1 (analysis from [22]). Comparison Z Punadjusted Padjusted Constant—Control 1.30 0.20 0.20 Constant—Varying -2.59 0.01 0.01 Control—Varying -3.89 0.0001 0.0003 Spread is calculated by the absolute value of deviation angles from each measurement with the mean vector for that condition. The statistical test used was the Kruskal-Wallis multiple comparison (using R program Kruskal.test; p- values adjusted with the Benjamini-Hochberg method [23] using R program dunnTest). The p-values indicate the probability that the two groups share the same spread of vanishing bearings. https://doi.org/10.1371/journal.pone.0241280.t003 in scatter between the constant field exposure and the control group was not statistically signif- icant (p = 0.20). Finally, no differences were observed in the total time it took each pigeon to enter the loft (p = 0.58). After the first release, one pigeon from the varying group failed to return to the loft; it was probably lost and/or killed by a predator. Two of the control birds arrived after the microchip reader was removed 72 hours after release; and one of the constant field birds was found at another loft and was returned to its home loft >72 hours after release; the arrival time for these birds was listed as 72 h for statistical analyses. Second release Fig 2 shows the vanishing bearings for birds during the second release, 23 days after exposure to all 3 conditions. No significant differences were detected between groups in the following parameters evaluated: difference in mean vanishing bearing between pairs of groups (see Table 2), dispersion of vanishing bearings with respect to the mean angle for each group (Kruskal-Wallace test across all 3 groups p = 0.30), time from release to vanishing as the pigeon headed back home (p = 0.40), mean duration from release to entering the loft (p = 0.60). How- ever, all three groups showed a significant difference between the mean vanishing bearing and the bearing to the loft (control p = 0.002; constant field p = 0.046; varying field p = 0.0005). The fastest pigeon in this case took 70 minutes to arrive and 26 out of the 29 arrived within the first 24h. One animal from the varying field group failed to return in this second release. Discussion After exposure to the varying magnetic field from standard MR acquisition sequences in a 3T MRI, the exposed pigeons showed a significant increase in the scatter of their vanishing bear- ings. The mean bearing for the group exposed to the varying magnetic field was farther from the correct direction angle of departure towards the pigeon loft than those of the control group and the group exposed to a constant field. The difference between the mean bearing for the varying field condition group and the home direction was not significant, but had a trend at the p = 0.07 level. The pigeons exposed to the constant magnetic field did not show any signifi- cant differences in behavior with respect to the controls in any of the response parameters that we measured. To our knowledge this is the first demonstration of deficits in orientation after exposure to magnetic fields used for clinical diagnostic purposes in animals that use geomagnetism for orientation. Despite the increased scatter in orientation of the pigeons exposed to the varying fields, they could overcome this adverse effect and all but one pigeon returned to the loft PLOS ONE | https://doi.org/10.1371/journal.pone.0241280 December 18, 2020 6/9 PLOS ONE 3T MR imaging affects orientation of homing pigeons Fig 2. Vanishing bearings of pigeons from the second release along with the mean vector for each group. Black = control; Orange = exposed to constant magnetic field in MRI; Red = exposed to varying magnetic field from scanning sequences. Home direction is marked with the light blue line and “H”. https://doi.org/10.1371/journal.pone.0241280.g002 within thirty five hours of release. Surprisingly, even though the pigeons exposed to the acqui- sition sequences came out more scattered, they entered the pigeon loft (average time interval from release to entering the loft) earlier. These latter results suggest that while exposure to the varying field affected orientation behavior immediately upon release, it did not have a strong effect on overall homing behavior. There was higher scatter and more deviation of the mean vector from the home bearing for birds in all three conditions during the second release compared to control and constant expo- sure birds in the first release, indicating a release site bias at the more distant second site. In fact, the topography of the second release site could explain the common angle of deviation in all three groups compared to the direct line to the loft. Pigeons typically tend not to fly over mountains or forest areas but try to maintain their routes through open spaces as much as pos- sible to minimize effort and potential risk from predators. As there were forested areas and some elevation of the terrain at approximately 1.2 km distance from the second release point in a straight line to the loft, animals in all three groups may have chosen to deviate consistently some degrees to the south to avoid those areas. PLOS ONE | https://doi.org/10.1371/journal.pone.0241280 December 18, 2020 7/9 PLOS ONE 3T MR imaging affects orientation of homing pigeons The present study reveals that diagnostic MR may not be innocuous for certain animal spe- cies. If the effect stems from interference with the mechanisms that some species use to sense geomagnetism, then the effect may be limited to species, such as homing pigeons, migratory birds or sea turtles, that are able to sense geomagnetic fields. If the effect of the exposure to MR sequences is not specific to the magnetic sense, then the adverse effect on orientation found here may also apply to other consequences in animals that do not use magnetic cues for orien- tation. This raises concerns that clinical evaluations using MRI may cause temporary afteref- fects that impair orientation and possibly other behavioral or physiological processes. Further research is needed in order to better discriminate the potential impact and the duration of del- eterious effects depending on the species and the different MR protocols used. Supporting information S1 Table. Response data from the 3 groups of pigeons from release site 1 and release site 2. (PDF) Acknowledgments We gratefully acknowledge assistance from Arthur Pewsey and Graeme D Ruxton with analy- ses from their book “Circular statistics in R.” Author Contributions Conceptualization: Daniel Garcı́a Párraga, Vicente Marco-Cabedo, Luis Martı́-Bonmatı́. Data curation: Daniel Garcı́a Párraga, Vicente Marco-Cabedo. Formal analysis: Peter L. Tyack, Luis Martı́-Bonmatı́. Funding acquisition: Daniel Garcı́a Párraga. Investigation: Vicente Marco-Cabedo, José Luis Crespo-Picazo. Methodology: Daniel Garcı́a Párraga, Vicente Marco-Cabedo, José Luis Crespo-Picazo, Luis Martı́-Bonmatı́. Project administration: Daniel Garcı́a Párraga, Vicente Marco-Cabedo, José Luis Crespo- Picazo. Resources: Daniel Garcı́a Párraga, Vicente Marco-Cabedo. Supervision: Daniel Garcı́a Párraga, Vicente Marco-Cabedo, José Luis Crespo-Picazo. Validation: Daniel Garcı́a Párraga, Vicente Marco-Cabedo. Visualization: Peter L. Tyack, Vicente Marco-Cabedo. Writing – original draft: Daniel Garcı́a Párraga, Vicente Marco-Cabedo, José Luis Crespo- Picazo. Writing – review & editing: Daniel Garcı́a Párraga, Peter L. Tyack, Vicente Marco-Cabedo, José Luis Crespo-Picazo, Xavier Manteca, Luis Martı́-Bonmatı́. References 1. Gould JL. Magnetic field sensitivity in animals. Annual Review of Physiology. 1984; 46(1): 585–98. https://doi.org/10.1146/annurev.ph.46.030184.003101 PMID: 6370118 PLOS ONE | https://doi.org/10.1371/journal.pone.0241280 December 18, 2020 8/9 PLOS ONE 3T MR imaging affects orientation of homing pigeons 2. Wiltschko W, Wiltschko R. Magnetic orientation and magnetoreception in birds and other animals. Jour- nal of Comparative Physiology A. 2005; 191(8):675–93. https://doi.org/10.1007/s00359-005-0627-7 PMID: 15886990 3. Marhold S, Wiltschko W, Burda H. A magnetic polarity compass for direction finding in a subterranean mammal. Naturwissenschaften. 1997; 84(9): 421–423. 4. Meyer CG, Holland KN, Papastamatiou YP. Sharks can detect changes in the geomagnetic field. Jour- nal of the Royal Society Interface. 2004; 2(2): 129–130. 5. Hsu C-Y, Ko F-Y, Li C-W, Fann K, Lue J-T. Magnetoreception system in honeybees (Apis mellifera). PloS One. 2007; 2(4): e395. https://doi.org/10.1371/journal.pone.0000395 PMID: 17460762 6. Lohmann KJ, Luschi P, Hays GC. Goal navigation and island-finding in sea turtles. Journal of Experi- mental Marine Biology and Ecology. 2008; 356(1–2): 83–95. 7. Vácha M, Půžová T, Kvı́ćalová M. Radio frequency magnetic fields disrupt magnetoreception in Ameri- can cockroach. Journal of Experimental Biology. 2009; 212(21): 3473–3477. https://doi.org/10.1242/ jeb.028670 PMID: 19837889 8. Wiltschko R, Thalau P, Gehring D, Niessner C, Ritz T, Wiltschko W. Magnetoreception in birds: the effect of radio-frequency fields. Journal of the Royal Society Interface. 2015; 12(20141103). https://doi. org/10.1098/rsif.2014.1103 PMID: 25540238 9. Steinmetz HW, Makara M. The use of computed tomography and magnetic resonance imaging in zoo animals. In Miller RE, Fowler ME, editors. Zoo and Wild Animal Medicine. St Louis: Elsevier Saunders; 2014. pp. 661–666. 10. Ahmed S, Shellock FG. Magnetic resonance imaging safety: implications for cardiovascular patients. Journal of Cardiovascular Magnetic Resonance. 2001; 3(3): 171–182. https://doi.org/10.1081/jcmr- 100107466 PMID: 11816614 11. Chakeres DW, de Vocht F. Static magnetic field effects on human subjects related to magnetic reso- nance imaging systems. Progress in Biophysics and Molecular Biology 2005; 87(2–3): 255–265. https:// doi.org/10.1016/j.pbiomolbio.2004.08.012 PMID: 15556664 12. de Vocht F, van Drooge H, Engels H, Kromhout H. Exposure, health complaints and cognitive perfor- mance among employees of an MRI scanners manufacturing department. Journal of Magnetic Reso- nance Imaging: An Official Journal of the International Society for Magnetic Resonance in Medicine. 2006; 23(2): 197–204. 13. Modo M, Bulte J. Magnetic Resonance Neuroimaging. Methods in Molecular Biology. Berlin: Springer Verlag; 2011. 14. Tsuji Y, Nakagawa M, Suzuki Y. Five-Tesla static magnetic fields suppress food and water consumption and weight gain in mice. Industrial Health 1996; 34(4): 347–357. https://doi.org/10.2486/indhealth.34. 347 PMID: 8908845 15. Strbak O, Kopcansky P, Frollo I. Biogenic magnetite in humans and new magnetic resonance hazard questions. Measurement Science Review. 2011; 11(3): 85–91. 16. Valente ALS, Cuenca R, Zamora MA, Parga ML, Lavin S, Alegre F, et al. Sectional anatomic and mag- netic resonance imaging features of coelomic structures of loggerhead sea turtles. American Journal of Veterinary Research. 2006; 67(8): 1347–1353. https://doi.org/10.2460/ajvr.67.8.1347 PMID: 16881846 17. Wyneken J. Computed Tomography and Magnetic Resonance Imaging. In: Mader DR, Divers SJ, edi- tors. Current Therapy in Reptile Medicine and Surgery. Amsterdam: Elsevier Health Sciences; 2013. pp. 93–106. 18. Wallraff HG. Avian navigation: pigeon homing as a paradigm. Berlin: Springer Science & Business Media; 2005. 19. Wiltschko W, Wiltschko R. Homing pigeons as a model for avian navigation? Journal of Avian Biology. 2017; 48(1): 66–74. 20. Pewsey A, Neuhäuser M, Ruxton GD. Circular statistics in R. Oxford: Oxford University Press; 2013. 21. Berens P. CircStat: A Matlab Toolbox for Circular Statistics. Journal of Statistical Software. 2009; 31 (10):1–21. 22. Wallraff HG. Goal-oriented and compass-oriented movements of displaced homing pigeons after con- finement in differentially shielded aviaries. Behavioral Ecology and Sociobiology. 1979; 5(2): 201–225. 23. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to mul- tiple testing. Journal of the Royal Statistical Society. Series B (Methodological). 1995; 57: 289–300. 24. Dunn OJ. Multiple comparisons using rank sums. Technometrics. 1964; 6(3): 241–252. PLOS ONE | https://doi.org/10.1371/journal.pone.0241280 December 18, 2020 9/9