USGS Science⌕ Search

SEARCH · USGS Science

Results for “Behavioral and Neural Biology”

Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

5 recordsLinked to original sources
USGS1980

Sleeping distance in wild wolf packs

Sleeping distances were observed among members of 13 wild wolf (Canis lupus) packs and 11 pairs in northeastern Minnesota to determine if the distances correlated with pack size and composition. The study utilized aerial radio-tracking and observation during winter. Pack size and number of adults per pack were inversely related to pack average sleeping distance and variability. No correlation between sleeping distance and microclimate was observed. Possible relationships between social bonding and our results are discussed.

Behavioral and Neural Biology↗
USGS1986

Effect of anosmia on reproduction in male and female wolves (Canis lupus)

Anosmia was produced in two female and three male wolves by transection of the olfactory peduncle and was confirmed by their inability to detect meat, urine, feces, anal-gland secretions, and fish emulsion. All operated animals continued to investigate the environment with their noses, to interact normally with other pack members, and to feed at levels which maintained presurgical body weights. No effect was found on reproductive physiology (females: estradiol or progesterone concentrations, ovulation, pregnancy or parturition; males: testosterone, testicular recrudescence or sperm numbers, motility or maturation). One anosmic female became dominant and although she urine-marked with a flexed leg, the rate was lower than typical for dominant females and perhaps contributed to her failure to pair-bond with the dominant male. One anosmic male raised-leg-urinated while competing for pack dominance and when kenneled away from other males. Precopulatory, copulatory, and maternal behavior were observed for one anosmic female and appeared normal. However, neither male that was sexually naive before surgery showed interest in proestrous or estrous females. The possibility that secondary degeneration of brain regions mediating sexual behavior was responsible for the failure of these males to respond was not supported. Not only was the lack of male sexual response the only serious deficit following transection, but the male which was sexually experienced prior to surgery did copulate successfully during his second postoperative breeding season despite continued anosmia. Chemosensory priming from female urine during the protracted proestrous phase, as well as urinary and vaginal odors during estrus, appear to be critical for induction of full sexual potency in sexually naive males. The importance of urine and vaginal secretions in the sexual response of experienced males is uncertain.

Behavioral and Neural Biology↗
USGS2025

Achieving interpretable machine learning by functional decomposition of black-box models into explainable predictor effects

Machine learning (ML) models are often based on complex black-box architectures that are difficult to interpret. This interpretability problem can hinder the use of ML in fields like medicine, ecology, and insurance, and has boosted research in interpretable machine learning (IML). Here, we propose a novel approach for the functional decomposition of black-box predictions, which is a core concept of IML. This approach replaces the prediction function with a surrogate model consisting of simpler subfunctions, providing insights into the direction and strength of the main feature contributions and their interactions. Our method is based on a concept termed “stacked orthogonality”, which ensures that the main effects capture as much functional behavior as possible. To compute the subfunctions, we combine neural additive modeling with an efficient post-hoc orthogonalization procedure. Our method yielded plausible results in an analysis of stream biological condition in the Chesapeake Bay watershed (United States).

Chesapeake Bay watershed↗
USGS2013

Comparing mechanisms of host manipulation across host and parasite taxa

Parasites affect host behavior in several ways. They can alter activity, microhabitats or both. For trophically transmitted parasites (the focus of our study), decreased activity might impair the ability of hosts to respond to final-host predators, and increased activity and altered microhabitat choice might increase contact rates between hosts and final-host predators. In an analysis of trophically transmitted parasites, more parasite groups altered activity than altered microhabitat choice. Parasites that infected vertebrates were more likely to impair the host’s reaction to predators, whereas parasites that infected invertebrates were more likely to increase the host’s contact with predators. The site of infection might affect how parasites manipulate their hosts. For instance, parasites in the central nervous system seem particularly suited to manipulating host behavior. Manipulative parasites commonly occupy the body cavity, muscles and central nervous systems of their hosts. Acanthocephalans in the data set differed from other taxa in that they occurred exclusively in the body cavity of invertebrates. In addition, they were more likely to alter microhabitat choice than activity. Parasites in the body cavity (across parasite types) were more likely to be associated with increased host contact with predators. Parasites can manipulate the host through energetic drain, but most parasites use more sophisticated means. For instance, parasites target four physiological systems that shape behavior in both invertebrates and vertebrates: neural, endocrine, neuromodulatory and immunomodulatory. The interconnections between these systems make it difficult to isolate specific mechanisms of host behavioral manipulation.

Journal of Experimental Biology↗
USGS2015

Hydrologic record extension of water-level data in the Everglades Depth Estimation Network (EDEN), 1991-99

The real-time Everglades Depth Estimation Network (EDEN) has been established to support a variety of scientific and water management purposes. The expansiveness of the Everglades, limited number of gaging stations, and extreme sensitivity of the ecosystem to small changes in water depth have created a need for accurate water-level and water-depth maps. The EDEN water-surface elevation model uses data from approximately 240 gages in the Everglades to create daily continuous interpolations of the water-surface elevation and water depth for the freshwater portion of the Everglades from 2000 to the present (2014). These maps provide hydrologic data previously unavailable for assessing biological and ecological studies. Ecologists working in the Everglades expressed a need to the EDEN project team for daily EDEN water-level surfaces from 1990 to 1999. The additional 10 years of surfaces will provide ecologists and resource managers with two decades (1991–2011) of surfaces to analyze hydrologic dynamics. Before 2000, many of the EDEN gages used to generate water surfaces were not in operation. These datasets were extended to provide estimations of hydrologic time-series histories. The general approach to the record extension (hindcasts) was to (1) create a database of available data from 1990 to the present; (2) use dynamic cluster analysis to group stations with similar hydrologic behaviors for subareas of the Everglades with a large number of stations; (3) use results from the cluster analysis to select candidate explanatory variables; (4) develop linear regression or artificial neural network models to extend water-level records; and (5) evaluate record extensions by using model performance statistics and comparison of water-surface maps for similar hydrologic conditions for the hindcasted period (1991–99) and measured period (2000–11). To hindcast and fill data records, 214 empirical models were developed—189 are linear regression models and 25 are artificial neural network models. The coefficient of determination (R 2 ) for 163 of the models is greater than 0.80 and the median percent model error (root mean square error divided by the range of the measured data) is 5 percent. To evaluate the performance of the hindcast models as a group, contour maps of modeled water-level surfaces at 2-centimeter (cm) intervals were generated using the hindcasted data. The 2-cm contour maps were examined for selected days to verify that water surfaces from the EDEN model are consistent with the input data. The biweekly 2-cm contour maps did show a higher number of issues during days in 1990 as compared to days after 1990. May 1990 had the lowest water levels in the Everglades of the 21-year dataset used for the hindcasting study. To hindcast these record low conditions in 1990, many of the hindcast models would require large extrapolations beyond the range of the predictive quality of the models. For these reasons, it was decided to limit the hindcasted data to the period January 1, 1991, to December 31, 1999. Overall, the hindcasted and gap-filled data are assumed to provide reasonable estimates of station-specific water-level data for an extended historical period to inform research and natural resource management in the Everglades.

Florida↗