Registered reports are a new publication workflow where the decision to publish is made prior to data collection and analysis and thus cannot be dependent on the outcome of the study. An increasing number of journals have adopted this new mechanism, but previous research suggests that submission rates are still relatively low. We conducted a census of journals publishing registered reports (N = 278) using independent coders to collect information from submission guidelines, with the goal of documenting journals’ early adoption of registered reports. Our results show that the majority of journals adopting registered reports are in psychology, and it typically takes about a year to publish the first registered report after adopting. Still, many journals have not published their first registered report. There is high variability in impact of journals adopting registered reports. Many journals do not include concrete information about policies that address concerns about registered reports (e.g., exploratory analysis); however, those that do typically allow these practices with some restrictions. Additionally, other open science practices are commonly encouraged or required as part of the registered report process, especially open data and materials. Overall, many journals did not include many of the fields coded by the research team,more »which could be a barrier to submission for some authors. Though the majority of journals allow authors to be anonymous during the review process, a sizable portion do not, which could also be a barrier to submission. We conclude with future directions and implications for authors of registered reports, journals that have already adopted registered reports, and journals that may consider adopting registered reports in the future.« less
Cephalopods’ remarkable behavior and complex neurobiology make them valuable comparative model organisms, but studies aimed at enhancing welfare of captive cephalopods remain uncommon. Increasing regulation of cephalopods in research laboratories has resulted in growing interest in welfare-oriented refinements, including analgesia and anesthesia. Although general and local anesthesia in cephalopods have received limited prior study, there have been no studies of systemic analgesics in cephalopods to date. Here we show that analgesics from several different drug classes may be effective in E. berryi. Buprenorphine, ketorolac and dexmedetomidine, at doses similar to those used in fish, showed promising effects on baseline nociceptive thresholds, excitability of peripheral sensory nerves, and on behavioral responses to transient noxious stimulation. We found no evidence of positive effects of acetaminophen or ketamine administered at doses that are effective in vertebrates. Bioinformatic analyses suggested conserved candidate receptors for dexmedetomidine and ketorolac, but not buprenorphine. We also show that rapid general immersion anesthesia using a mix of MgCl2 and ethanol was successful in E. berryi at multiple age classes, similar to findings in other cephalopods. These data indicate that systemic analgesia and general anesthesia in Euprymna berryi are achievable welfare enhancing interventions, but further study and refinement is warranted.
Buckwalter, Grace
Chhin(
, Proceedings of the IEEE Signal Processing in Medicine and Biology Symposium (SPMB))
Obeid, Iyad
Selesnick
(Ed.)
The Temple University Hospital EEG Corpus (TUEG) [1] is the largest publicly available EEG corpus of its type and currently has over 5,000 subscribers (we currently average 35 new subscribers a week). Several valuable subsets of this corpus have been developed including the Temple University Hospital EEG Seizure Corpus (TUSZ) [2] and the Temple University Hospital EEG Artifact Corpus (TUAR) [3]. TUSZ contains manually annotated seizure events and has been widely used to develop seizure detection and prediction technology [4]. TUAR contains manually annotated artifacts and has been used to improve machine learning performance on seizure detection tasks [5]. In this poster, we will discuss recent improvements made to both corpora that are creating opportunities to improve machine learning performance. Two major concerns that were raised when v1.5.2 of TUSZ was released for the Neureka 2020 Epilepsy Challenge were: (1) the subjects contained in the training, development (validation) and blind evaluation sets were not mutually exclusive, and (2) high frequency seizures were not accurately annotated in all files. Regarding (1), there were 50 subjects in dev, 50 subjects in eval, and 592 subjects in train. There was one subject common to dev and eval, five subjects common to dev andmore »train, and 13 subjects common between eval and train. Though this does not substantially influence performance for the current generation of technology, it could be a problem down the line as technology improves. Therefore, we have rebuilt the partitions of the data so that this overlap was removed. This required augmenting the evaluation and development data sets with new subjects that had not been previously annotated so that the size of these subsets remained approximately the same. Since these annotations were done by a new group of annotators, special care was taken to make sure the new annotators followed the same practices as the previous generations of annotators. Part of our quality control process was to have the new annotators review all previous annotations. This rigorous training coupled with a strict quality control process where annotators review a significant amount of each other’s work ensured that there is high interrater agreement between the two groups (kappa statistic greater than 0.8) [6]. In the process of reviewing this data, we also decided to split long files into a series of smaller segments to facilitate processing of the data. Some subscribers found it difficult to process long files using Python code, which tends to be very memory intensive. We also found it inefficient to manipulate these long files in our annotation tool. In this release, the maximum duration of any single file is limited to 60 mins. This increased the number of edf files in the dev set from 1012 to 1832. Regarding (2), as part of discussions of several issues raised by a few subscribers, we discovered some files only had low frequency epileptiform events annotated (defined as events that ranged in frequency from 2.5 Hz to 3 Hz), while others had events annotated that contained significant frequency content above 3 Hz. Though there were not many files that had this type of activity, it was enough of a concern to necessitate reviewing the entire corpus. An example of an epileptiform seizure event with frequency content higher than 3 Hz is shown in Figure 1. Annotating these additional events slightly increased the number of seizure events. In v1.5.2, there were 673 seizures, while in v1.5.3 there are 1239 events. One of the fertile areas for technology improvements is artifact reduction. Artifacts and slowing constitute the two major error modalities in seizure detection [3]. This was a major reason we developed TUAR. It can be used to evaluate artifact detection and suppression technology as well as multimodal background models that explicitly model artifacts. An issue with TUAR was the practicality of the annotation tags used when there are multiple simultaneous events. An example of such an event is shown in Figure 2. In this section of the file, there is an overlap of eye movement, electrode artifact, and muscle artifact events. We previously annotated such events using a convention that included annotating background along with any artifact that is present. The artifacts present would either be annotated with a single tag (e.g., MUSC) or a coupled artifact tag (e.g., MUSC+ELEC). When multiple channels have background, the tags become crowded and difficult to identify. This is one reason we now support a hierarchical annotation format using XML – annotations can be arbitrarily complex and support overlaps in time. Our annotators also reviewed specific eye movement artifacts (e.g., eye flutter, eyeblinks). Eye movements are often mistaken as seizures due to their similar morphology [7][8]. We have improved our understanding of ocular events and it has allowed us to annotate artifacts in the corpus more carefully. In this poster, we will present statistics on the newest releases of these corpora and discuss the impact these improvements have had on machine learning research. We will compare TUSZ v1.5.3 and TUAR v2.0.0 with previous versions of these corpora. We will release v1.5.3 of TUSZ and v2.0.0 of TUAR in Fall 2021 prior to the symposium. ACKNOWLEDGMENTS Research reported in this publication was most recently supported by the National Science Foundation’s Industrial Innovation and Partnerships (IIP) Research Experience for Undergraduates award number 1827565. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the official views of any of these organizations. REFERENCES [1] I. Obeid and J. Picone, “The Temple University Hospital EEG Data Corpus,” in Augmentation of Brain Function: Facts, Fiction and Controversy. Volume I: Brain-Machine Interfaces, 1st ed., vol. 10, M. A. Lebedev, Ed. Lausanne, Switzerland: Frontiers Media S.A., 2016, pp. 394 398. https://doi.org/10.3389/fnins.2016.00196. [2] V. Shah et al., “The Temple University Hospital Seizure Detection Corpus,” Frontiers in Neuroinformatics, vol. 12, pp. 1–6, 2018. https://doi.org/10.3389/fninf.2018.00083. [3] A. Hamid et, al., “The Temple University Artifact Corpus: An Annotated Corpus of EEG Artifacts.” in Proceedings of the IEEE Signal Processing in Medicine and Biology Symposium (SPMB), 2020, pp. 1-3. https://ieeexplore.ieee.org/document/9353647. [4] Y. Roy, R. Iskander, and J. Picone, “The NeurekaTM 2020 Epilepsy Challenge,” NeuroTechX, 2020. [Online]. Available: https://neureka-challenge.com/. [Accessed: 01-Dec-2021]. [5] S. Rahman, A. Hamid, D. Ochal, I. Obeid, and J. Picone, “Improving the Quality of the TUSZ Corpus,” in Proceedings of the IEEE Signal Processing in Medicine and Biology Symposium (SPMB), 2020, pp. 1–5. https://ieeexplore.ieee.org/document/9353635. [6] V. Shah, E. von Weltin, T. Ahsan, I. Obeid, and J. Picone, “On the Use of Non-Experts for Generation of High-Quality Annotations of Seizure Events,” Available: https://www.isip.picone press.com/publications/unpublished/journals/2019/elsevier_cn/ira. [Accessed: 01-Dec-2021]. [7] D. Ochal, S. Rahman, S. Ferrell, T. Elseify, I. Obeid, and J. Picone, “The Temple University Hospital EEG Corpus: Annotation Guidelines,” Philadelphia, Pennsylvania, USA, 2020. https://www.isip.piconepress.com/publications/reports/2020/tuh_eeg/annotations/. [8] D. Strayhorn, “The Atlas of Adult Electroencephalography,” EEG Atlas Online, 2014. [Online]. Availabl« less
Climate change is affecting the Arctic at an unprecedented rate, potentially releasing substantial amounts of greenhouse gases (CO2 (carbon dioxide) and CH4 (Methane)) from tundra ecosystems. Measuring greenhouse gas emissions in the Arctic, particularly outside of the summer period, is very challenging due to extreme weather conditions. This research project provided the first annual balance of both CH4 and CO2 fluxes in a total of five sites spanning a 300Km transect across the North Slope of Alaska (three sites in Barrow, one site in Aquasuk, and one site in Ivotuk). The results from the continuous year-round CH4 fluxes across these sites showed how cumulative emissions for the cold season accounted on average for 50% of the annual budget (Zona et al., 2016), a notably higher contribution than previously modelled, and also higher than observed in boreal Alaska. The analysis of the cold period CH4 fluxes suggested that the presence of an unfrozen soil layer in the fall and early winter was a major control on cold season CH4 emissions (Zona et al., 2016). We also cross-compared all instruments measuring ecosystem scale CO2 and CH4 fluxes operating at our sites, which allowed us to make recommendation of the best performing
instruments under these extreme weather conditions. The best performing instruments were closed path analyzers and intermittently heated sonic anemometers which had the highest final data cover. A continuously heated anemometer increased data coverage relative to non-heated anemometers, but resulted in an overestimation of the fluxes (Goodrich et al., 2016). We developed an intermittent heating strategy that was only activated when the data quality was low, and appeared to be the preferable method to prevent icing while avoiding biases to the measurements. Closed and open-path analyzers showed good agreement, but data coverage was much greater when using closed-path analyzers, especially during winter (Goodrich et al., 2016). Given the importance of vegetation on greenhouse gas emissions, we also investigated the role of different vegetation types under a broad range of environmental conditions on the CH4 emissions. We found that vegetation type can be a very useful tool to describe the spatial variability in CH4 emissions over the landscape (McEwing et al., 2015), and that just two vegetation types were able to explain about 50% of the variability in CH4 fluxes across ecosystems even hundreds of kilometers apart (Davidson et al., 2016a). To upscale these plot scale fluxes we completed high resolution vegetation maps in each of our tower sites (Davidson et al., 2016b), which are the finest resolution maps currently available from these sites, and also contributed to larger scale mapping effort (Walker et al., 2016). The soil microbial analysis from soil cores collected across our sites showed an association between overall microbial diversity and latitude, with a higher diversity found in the northerly site and lower diversity in the southerly site, contrary to current knowledge (Wagner et al., accepted). We also measured CH4 and CO2 concentrations in the soil, which showed to be orders of magnitude higher than in the atmosphere (Arndt et al., 2016). Our results contributed to model development (Xu et al., 2016; Kobayashi et al., 2016; Liljedahl et al., 2016; Luus et al., 2017), and to a wide variety of other projects as shown by the hundreds of download of our data from Ameriflux. Overall, this grant resulted in the publication of 25 peer reviewed journal articles, including in high impact factor journals such as PNAS (Proceedings of the National Academy of Sciences of the United States of America), and Nature Climate Change, in addition to five more in review and in preparation, and supported the research of seven PhD students, two master students, and ten undergraduate students. More>>
Rasys, Ashley M; Divers, Stephen J; Lauderdale, James D; Menke, Douglas B(
, Laboratory Animals)
Anolis lizards have served as important research models in fields ranging from evolution and ecology to physiology and biomechanics. However, anoles are also emerging as important models for studies of embryo development and tissue regeneration. The increased use of anoles in the laboratory has produced a need to establish effective methods of anesthesia, both for routine veterinary procedures and for research procedures. Therefore, we tested the efficacy of different anesthetic treatments in adult female Anolis sagrei. Alfaxalone, dexmedetomidine, hydromorphone, ketamine and tribromoethanol were administered subcutaneously (SC), either alone or combined at varying doses in a total of 64 female anoles. Drug induction time, duration, anesthesia level and adverse effects were assessed. Differences in anesthesia level were observed depending on injection site and drug combination. Alfaxalone/dexmedetomidine and tribromoethanol/dexmedetomidine were the most effective drug combinations for inducing a surgical plane of anesthesia in anoles. Brown anoles injected SC with alfaxalone (30 mg/kg) plus dexmedetomidine (0.1 mg/kg) or with tribromoethanol (400 mg/kg) plus dexmedetomidine (0.1 mg/kg) experienced mean durations of surgical anesthesia levels of 31.2 ± 5.3 and 87.5 ± 19.8 min with full recovery after another 10.9 ± 2.9 and 46.2 ± 41.8 min, respectively. Hydromorphone given with alfaxalone/dexmedetomidine resulted in deep anesthesia with respiratory depression, while ketamine/hydromorphone/dexmedetomidine produced only light to moderate sedation. We determinedmore »that alfaxalone/dexmedetomidine or tribromoethanol/dexmedetomidine combinations were sufficient to maintain a lizard under general anesthesia for coeliotomy. This study represents a significant step towards understanding the effects of anesthetic agents in anole lizards and will benefit both veterinary care and research on these animals.« less
Larry Carbone, Jamie Austin. Pain and Laboratory Animals: Publication Practices for Better Data Reproducibility and Better Animal Welfare. Retrieved from https://par.nsf.gov/biblio/10024804. PloS one 11.5 Web. doi:10.1371/journal. pone.0155001.
Larry Carbone, Jamie Austin. Pain and Laboratory Animals: Publication Practices for Better Data Reproducibility and Better Animal Welfare. PloS one, 11 (5). Retrieved from https://par.nsf.gov/biblio/10024804. https://doi.org/10.1371/journal. pone.0155001
