Super-resolution neural networks improve the spatiotemporal resolution of adaptive MRI-guided radiation therapy.

BACKGROUND: Magnetic resonance imaging (MRI) offers superb non-invasive, soft tissue imaging of the human body. However, extensive data sampling requirements severely restrict the spatiotemporal resolution achievable with MRI. This limits the modality's utility in real-time guidance applications, particularly for the rapidly growing MRI-guided radiation therapy approach to cancer treatment. Recent advances in artificial intelligence (AI) could reduce the trade-off between the spatial and the temporal resolution of MRI, thus increasing the clinical utility of the imaging modality. METHODS: We trained deep learning-based super-resolution neural networks to increase the spatial resolution of real-time MRI. We developed a framework to integrate neural networks directly onto a 1.0 T MRI-linac enabling real-time super-resolution imaging. We integrated this framework with the targeting system of the MRI-linac to demonstrate real-time beam adaptation with super-resolution-based imaging. We tested the integrated system using large publicly available datasets, healthy volunteer imaging, phantom imaging, and beam tracking experiments using bicubic interpolation as a baseline comparison. RESULTS: Deep learning-based super-resolution increases the spatial resolution of real-time MRI across a variety of experiments, offering measured performance benefits compared to bicubic interpolation. The temporal resolution is not compromised as measured by a real-time adaptation latency experiment. These two effects, an increase in the spatial resolution with a negligible decrease in the temporal resolution, leads to a net increase in the spatiotemporal resolution. CONCLUSIONS: Deployed super-resolution neural networks can increase the spatiotemporal resolution of real-time MRI. This has applications to domains such as MRI-guided radiation therapy and interventional procedures.

Magnetic resonance imaging (MRI) is a medical imaging modality that is used to image organs such as the brain, lungs, and liver as well as diseases such as cancer. MRI scans taken at high resolution are of overly long duration. This time constraint limits the accuracy of MRI-guided cancer radiation therapy, where imaging must be fast to adapt treatment to tumour motion. Here, we deployed artificial intelligence (AI) models to achieve fast and high detail MRI. We additionally validated our AI models across various scenarios. These AI-based models could potentially enable people with cancer to be treated with higher accuracy and precision.

Experimental comparison of linear regression and LSTM motion prediction models for MLC-tracking on an MRI-linac.

BACKGROUND: Magnetic resonance imaging (MRI)-guided radiotherapy with multileaf collimator (MLC)-tracking is a promising technique for intra-fractional motion management, achieving high dose conformality without prolonging treatment times. To improve beam-target alignment, the geometric error due to system latency should be reduced by using temporal prediction. PURPOSE: To experimentally compare linear regression (LR) and long-short-term memory (LSTM) motion prediction models for MLC-tracking on an MRI-linac using multiple patient-derived traces with different complexities. METHODS: Experiments were performed on a prototype 1.0 T MRI-linac capable of MLC-tracking. A motion phantom was programmed to move a target in superior-inferior (SI) direction according to eight lung cancer patient respiratory motion traces. Target centroid positions were localized from sagittal 2D cine MRIs acquired at 4 Hz using a template matching algorithm. The centroid positions were input to one of four motion prediction models. We used (1) a LSTM network which had been optimized in a previous study on patient data from another cohort (offline LSTM). We also used (2) the same LSTM model as a starting point for continuous re-optimization of its weights during the experiment based on recent motion (offline+online LSTM). Furthermore, we implemented (3) a continuously updated LR model, which was solely based on recent motion (online LR). Finally, we used (4) the last available target centroid without any changes as a baseline (no-predictor). The predictions of the models were used to shift the MLC aperture in real-time. An electronic portal imaging device (EPID) was used to visualize the target and MLC aperture during the experiments. Based on the EPID frames, the root-mean-square error (RMSE) between the target and the MLC aperture positions was used to assess the performance of the different motion predictors. Each combination of motion trace and prediction model was repeated twice to test stability, for a total of 64 experiments. RESULTS: The end-to-end latency of the system was measured to be (389 ± 15) ms and was successfully mitigated by both LR and LSTM models. The offline+online LSTM was found to outperform the other models for all investigated motion traces. It obtained a median RMSE over all traces of (2.8 ± 1.3) mm, compared to the (3.2 ± 1.9) mm of the offline LSTM, the (3.3 ± 1.4) mm of the online LR and the (4.4 ± 2.4) mm when using the no-predictor. According to statistical tests, differences were significant (p-value <0.05) among all models in a pair-wise comparison, but for the offline LSTM and online LR pair. The offline+online LSTM was found to be more reproducible than the offline LSTM and the online LR with a maximum deviation in RMSE between two measurements of 10%. CONCLUSIONS: This study represents the first experimental comparison of different prediction models for MRI-guided MLC-tracking using several patient-derived respiratory motion traces. We have shown that among the investigated models, continuously re-optimized LSTM networks are the most promising to account for the end-to-end system latency in MRI-guided radiotherapy with MLC-tracking.

Investigating the use of machine learning to generate ventilation images from CT scans.

BACKGROUND: Radiotherapy treatment planning incorporating ventilation imaging can reduce the incidence of radiation-induced lung injury. The gold-standard of ventilation imaging, using nuclear medicine, has limitations with respect to availability and cost. PURPOSE: An alternative type of ventilation imaging to nuclear medicine uses 4DCT (or breath-hold CT [BHCT] pair) with deformable image registration (DIR) and a ventilation metric to produce a CT ventilation image (CTVI). The purpose of this study is to investigate the application of machine learning as an alternative to DIR-based methods when producing CTVIs. METHODS: A patient dataset of 15 inhale and exhale BHCTs and Galligas PET ventilation images were used to train and test a 2D U-Net style convolutional neural network. The neural network established relationships between axial input BHCT image pairs and axial labeled Galligas PET images and was evaluated using eightfold cross-validation. Once trained, the neural network could produce a CTVI from an input BHCT image pair. The CTVIs produced by the neural network were qualitatively assessed visually and quantitatively compared to a Galligas PET ventilation image using a Spearman correlation and Dice similarity coefficient (DSC). The DSC measured the spatial overlap between three segmented equal lung volumes by ventilation (high, medium, and low functioning lung [LFL]). RESULTS: The mean Spearman correlation between the CTVIs and the Galligas PET ventilation images was 0.58 ± 0.14. The mean DSC over high, medium, and LFL between the CTVIs and Galligas PET ventilation images was 0.55 ± 0.06. Visually, a systematic overprediction of ventilation within the lung was observed in the CTVIs with respect to the Galligas PET ventilation images, with jagged regions of ventilation in the sagittal and coronal planes. CONCLUSIONS: A convolutional neural network was developed that could produce a CTVI from a BHCT image pair, which was then compared with a Galligas PET ventilation image. The performance of this machine learning method was comparable to previous benchmark studies investigating a DIR-based CTVI, warranting future development, and investigation of applying machine learning to a CTVI.

Workplace discrimination for persons with hearing loss: Before and after the 2008 ADA Amendments Act.

BACKGROUND: Individuals with hearing loss experience unique barriers to employment frequently documented in the areas of communication and education. The purpose of this article is to contribute to extend this inquiry to the uniqueness of workplace discrimination involving persons with hearing loss. OBJECTIVE: This study investigated differences in allegations of workplace discrimination filed by persons with hearing loss ("Hearing") compared to those filed by persons with other physical or neurological disabilities (General Disability, or "GENDIS") before and after the enactment of the 2008 Americans with Disabilities Act Amendments Act (2008 Amendments). METHODS: Using secondary data collected from the Equal Employment Opportunity Commission (EEOC) Integrated Mission System, we employ simple measures of proportion and odds ratios to describe differences between allegations derived from GENDIS and Hearing loss populations. These are population statistics, and not samples, of all allegations of discrimination reported to the EEOC through 2016. The comparisons involve Characteristics of the Charging Parties, Issues or discriminatory behaviors alleged, and closure statuses or Merit Rate of the EEOC's investigations - both before and after the 2008 Amendments. RESULTS: Following the 2008 Amendments, Charging Parties changed dramatically on age and gender status. Reasonable Accommodation, Hiring, Harassment, and employment Terms and Conditions showed unique features between groups and/or time periods. The "veracity" (confirmed truthfulness or merit) of the EEOC allegation (or Merit) rate also changed following the Amendments: higher for GENDIS; lower for Hearing. CONCLUSIONS: Possible rationale for these findings are offered, and new research questions are raised. Finally, implications for the cross-disability movement are presented.

Parasitic plasmid-host dynamics and host competition in flowing habitats.

Competition and coexistence were examined for two bacterial species, each potentially carrying a fitness-reducing, parasitic plasmid that was vertically transmitted with possible loss through segregation. Here, the fitness reduction of hosts was due to a toxin produced by plasmid-bearing cells and inhibiting plasmid-free cells. These populations were placed in a flow reactor habitat representing an idealized mammal gut. It was numerically shown that parasitic plasmids can mediate coexistence of competing host species, in conditions where plasmid-free hosts could not coexist. Numerical construction of a coexistence example suggests that it arises only for a narrow parameter range. In particular, both rates of segregation and the growth costs of plasmid carriage must be relatively low.

Sink or swim? Vertical movement and nutrient storage in phytoplankton.

A simulation model of vertically migrating phytoplankton is presented, using a Lagrangian, individual-based computational approach. Algal cells acquire and store nutrient at the bottom of the habitat, using stored nutrient to grow while in shallower waters. Stored nutrient also governs movement: cells sink when their nutrient quota falls below a threshold; otherwise they rise (or at least sink more slowly). Although the bottom of the habitat provides the growth-limiting nutrient, it also entails a risk of mortality. For a parameter set representing phosphorus-limited algae with a fixed nutrient storage capacity, neither continual sinking nor continual rising are optimal strategies. Instead, an adaptive dynamics approach suggests there is an optimal movement strategy in which cells rise when their storage capacity is partially filled, and otherwise sink. When the movement strategy is fixed in such a way and storage capacity is free to evolve, storage capacity approaches an optimal value several times higher than the minimal quota permitting population growth. Vertical movement and nutrient storage affect the vertical distribution of total nutrient. When cells always sink, total nutrient declines exponentially from the nutrient source at the bottom to a surface minimum. When cells always rise, there is a peak of total nutrient at the bottom, and another at the surface, with a minimum between. When cells move optimally, the vertical distribution of total nutrient can be close to uniform, or have a peak at mid-depth.

Complete Genome of Salmonella enterica Serovar Typhimurium T5-Like Siphophage Stitch.

Salmonellosis, caused by Salmonella, is a leading cause of food poisoning worldwide. With the continuing rise of bacterial antibiotic resistance, efforts are focused on seeking new approaches for treatment of bacterial infections, namely, bacteriophage therapy. Here, we report the complete genome of S. Typhimurium siphophage Stitch.

An exploratory study of air emissions associated with shale gas development and production in the Barnett Shale.

UNLABELLED: Information regarding air emissions from shale gas extraction and production is critically important given production is occurring in highly urbanized areas across the United States. Objectives of this exploratory study were to collect ambient air samples in residential areas within 61 m (200 feet) of shale gas extraction/production and determine whether a "fingerprint" of chemicals can be associated with shale gas activity. Statistical analyses correlating fingerprint chemicals with methane, equipment, and processes of extraction/production were performed. Ambient air sampling in residential areas of shale gas extraction and production was conducted at six counties in the Dallas/Fort Worth (DFW) Metroplex from 2008 to 2010. The 39 locations tested were identified by clients that requested monitoring. Seven sites were sampled on 2 days (typically months later in another season), and two sites were sampled on 3 days, resulting in 50 sets of monitoring data. Twenty-four-hour passive samples were collected using summa canisters. Gas chromatography/mass spectrometer analysis was used to identify organic compounds present. Methane was present in concentrations above laboratory detection limits in 49 out of 50 sampling data sets. Most of the areas investigated had atmospheric methane concentrations considerably higher than reported urban background concentrations (1.8-2.0 ppm(v)). Other chemical constituents were found to be correlated with presence of methane. A principal components analysis (PCA) identified multivariate patterns of concentrations that potentially constitute signatures of emissions from different phases of operation at natural gas sites. The first factor identified through the PCA proved most informative. Extreme negative values were strongly and statistically associated with the presence of compressors at sample sites. The seven chemicals strongly associated with this factor (o-xylene, ethylbenzene, 1,2,4-trimethylbenzene, m- and p-xylene, 1,3,5-trimethylbenzene, toluene, and benzene) thus constitute a potential fingerprint of emissions associated with compression. IMPLICATIONS: Information regarding air emissions from shale gas development and production is critically important given production is now occurring in highly urbanized areas across the United States. Methane, the primary shale gas constituent, contributes substantially to climate change; other natural gas constituents are known to have adverse health effects. This study goes beyond previous Barnett Shale field studies by encompassing a wider variety of production equipment (wells, tanks, compressors, and separators) and a wider geographical region. The principal components analysis, unique to this study, provides valuable information regarding the ability to anticipate associated shale gas chemical constituents.

Competition and allelopathy with resource storage: two resources.

Allelopathy is added to a familiar mathematical model of competition between two species for two essential resources in a chemostat environment. Both species store the resources, and each produces a toxin that induces mortality in the other species. The corresponding model without toxins displays outcomes of competitive exclusion independent of initial conditions, competitive exclusion that depends on initial conditions (bistability), and globally stable coexistence, depending on tradeoffs between competitors in growth requirements and consumption of the resources. Introducing toxins that act only between, and not within species, can destabilize coexistence leading to bistability or other multiple attractors. Invasibility of the missing species into a resident׳s semitrivial equilibrium is related to competitive outcomes. Mutual invasibility is necessary and sufficient for a globally stable coexistence equilibrium, but is not necessary for coexistence at a locally stable equilibrium. Invasibility of one semitrivial equilibrium but not the other is necessary but not sufficient for competitive exclusion independent of initial conditions. Mutual non-invasibility is necessary but not sufficient for bistability. Numerical analysis suggests that when competitors display bistability in the absence of toxin production, increases in the overall magnitude of resource supply cause bistability to arise over a larger range of supply ratios between the two resources. When competitors display coexistence in the absence of toxin production, increases in overall resource supply destabilize coexistence and produce bistability or other configurations of multiple attractors over large ranges of supply ratios. The emergence of multiple attractors at high resource supplies suggests that blooms of harmful algae producing allelopathic toxins could be difficult to predict under such rich conditions.

Ammonium treatments to suppress toxic blooms of Prymnesium parvum in a subtropical lake of semi-arid climate: results from in situ mesocosm experiments.

Prymnesium parvum is a haptophyte alga that forms toxic, fish-killing blooms in a variety of brackish coastal and inland waters. Its abundance and toxicity are suppressed by ammonium additions in laboratory cultures and aquaculture ponds. In a cove of a large reservoir (Lake Granbury, Texas, USA) with recurring, seasonal blooms of P. parvum, ammonium additions were tested in mesocosm enclosures for their ability to suppress blooms and their effects on non-target planktonic organisms. One experiment occurred prior to the peak abundance of a P. parvum bloom in the cove, and one encompassed the peak abundance and decline of the bloom. During 21-day experiments, weekly doses raised ammonium concentrations by either 10 or 40 µM. The added ammonium accumulated in experimental mesocosms, with little uptake by biota or other losses. Effects of ammonium additions generally increased over the course of the experiments. The higher ammonium dose suppressed the abundance and toxicity of P. parvum. The biomass of non-haptophyte algae was stimulated by ammonium additions, while positive, negative and neutral effects on zooplankton taxa were observed. Low ammonium additions insufficient to control P. parvum exacerbated its harmful effects. Our results indicate a potential for mitigating blooms of P. parvum with sufficient additions of ammonium to coves of larger lakes. However, factors excluded from mesocosms, such as dilution of ammonium by water exchange and sediment ammonium uptake, could reduce the effectiveness of such additions, and they would entail a risk of eutrophication from the added nitrogen.

Competition for one nutrient with internal storage and toxin mortality.

This study presents a mathematical model of two species competing in a chemostat for one resource that is stored internally, and who also compete through allelopathy. Each species produces a toxin to that increases mortality rate of its competitor. The two species system and its single species subsystem follow mass conservation constraints characteristic of chemostat models. Persistence of a single species occurs if the nutrient supply of an empty habitat allows it to acquire a threshold of stored nutrient quota, sufficient to overcome loss to outflow after accounting for the cost of toxin production. For the two-species system, a semitrivial equilibrium with one species resident is unstable to invasion by the missing species according to a similar threshold condition. The invader increases if acquires a stored nutrient quota sufficient to overcome loss to outflow and toxin-induced mortality, after accounting for the cost of the invader's own toxin production. If both semitrivial equilibria for the two-species system are invasible then there is at least one coexistence equilibrium. Numerical analyses indicate another possibility: bistability in which both semitrivial equilibria are stable against invasion. In such a case there is competitive exclusion of one species, whose identity depends on initial conditions. When there is a tradeoff between abilities to compete for the nutrient and to compete through toxicity, the more toxic species can dominate only under nutrient-rich conditions. Bistability under such conditions could contribute to the unpredictability of toxic algal blooms.

Competition between microorganisms for a single limiting resource with cell quota structure and spatial variation.

Microbial populations compete for nutrient resources, and the simplest mathematical models of competition neglect differences in the nutrient content of individuals. The simplest models also assume a spatially uniform habitat. Here both of these assumptions are relaxed. Nutrient content of individuals is assumed proportional to cell size, which varies for populations that reproduce by division, and the habitat is taken to be an unstirred chemostat where organisms and nutrients move by simple diffusion. In a spatially uniform habitat, the size-structured model predicts competitive exclusion, such that only the species with lowest break-even concentration persists. In the unstirred chemostat, coexistence of two competitors is possible, if one has a lower break-even concentration and the other can grow more rapidly. In all habitats, the calculation of competitive outcomes depends on a principal eigenvalue that summarizes relationships among cell growth, cell division, and cell size.

STIMULATING EFFECT OF ANABAENA SP. (CYANOBACTERIA) EXUDATE ON PRYMNESIUM PARVUM (HAPTOPHYTA)(1).

Prymnesium parvum blooms have become more frequent in the south-central United States, leading to significant ecological and economic impacts. Allelopathic effects from cyanobacteria were suggested as a mechanism that might limit the development of P. parvum blooms. This research focused on the effects of cultured cyanobacteria, Anabaena sp., on P. parvum. Over a 6-d period, daily additions of filtrate from the senescent Anabaena culture were made to P. parvum cultures growing in log phase. All treatments, including several types of controls, showed reductions in P. parvum biomass over the course of the experiment, but the treatments receiving Anabaena filtrate were reduced to a lesser degree, suggesting that filtrate from the senescent cyanobacteria culture was beneficial to P. parvum in some way. This unexpected outcome may have resulted from stimulation of heterotrophic bacteria by the addition of Anabaena filtrate, which likely contained exudates rich in dissolved organic carbon compounds. P. parvum was then able to supplement its nutritional requirements for growth by feeding on the elevated bacteria population. These findings coupled to previous observations suggest that interactions between cyanobacteria and P. parvum in natural environments are complex, where both allelopathic and growth-stimulating interactions are possible.

Resource storage and competition with spatial and temporal variation in resource availability.

This study addresses interspecific competition for a nutrient resource that is stored within individuals in habitats with both temporal and spatial variation. In such environments, population structure is induced by the mixture at any location of individuals with different amounts of stored nutrient, acquired elsewhere in the habitat. Focusing on phytoplankton competing for phosphorus in a partially mixed water column, an individual-based Lagrangian model is used to represent this population structure, and partial differential equations that approximate competitive dynamics are constructed by averaging over this population structure. Although the approximation model overestimates the benefit of resource storage to competitive fitness, both approaches predict that species with high storage capacity are favored by periodic resource pulses that are short lived but large in magnitude. Such storage specialists can competitively exclude or coexist with species that have advantages in maximal nutrient uptake and population growth rates. For very infrequent resource pulses, competitive dynamics become close to neutral. Thus, persistence of diverse species that are differentiated in nutrient storage and uptake capabilities is favored by resource pulses occurring with periods that are many times the average generation time of competitors.

Element content of Ochromonas danica: a replicated chemostat study controlling the growth rate and temperature.

Ecological stoichiometry focuses on the balance between multiple nutrient elements in resources and in consumers of those resources. The major consumers of bacteria in aquatic food webs are heterotrophic and mixotrophic nanoflagellates. Despite the importance of this consumer-resource interaction to understanding nutrient dynamics in the aquatic food web, few data are available addressing the element stoichiometry of flagellate consumers. Ochromonas danica, a mixotrophic bacterivore, was used as a model organism to study the relationships among temperature, growth rate and element stoichiometry. Ochromonas danica was grown in chemostats at dilution rates ranging between 0.03 and 0.10 h(-1) and temperatures ranging between 15 and 28 °C. Cells accumulated elements as interactive functions of temperature and growth rate, with the highest element concentrations corresponding to cells grown at a low temperature and high growth rates. The highest concentrations of elements were associated with small cells. Temperature and growth rate affected the element stoichiometry (as C:N, C:P and N:P) of O. danica in a complex manner, but the growth rate had a greater effect on ratios than did temperature.

Element stoichiometry of a mixotrophic protist grown under varying resource conditions.

The balance of essential elements (e.g. carbon [C], nitrogen [N], and phosphorus [P]) between consumers and their resources influences not only the growth and reproduction of the consumers but also the nutrients they regenerate. Flagellate protists are significant predators of aquatic bacteria and directly influence nutrient flow to higher trophic levels and, through excretion, influence the mineral element composition of dissolved nutrients. Because the element stoichiometry of protists is poorly characterized, we varied the resource composition of the bacterium Pseudomonas fluorescens and used it to grow the mixotrophic bacterivorous flagellate Ochromonas danica. Using a mass balance approach, the element composition of O. danica was found to vary depending upon the nutrient composition of the prey and ranged between 482:36:1 and 80:12:1 (C:N:P molar). Homeostasis plots suggested that flagellate protists weakly regulate their element composition and are likely to regenerate different elements depending upon the nature of the element limiting growth of their prey.

Coexistence of mixotrophs, autotrophs, and heterotrophs in planktonic microbial communities.

We examine what circumstances allow the coexistence of microorganisms following different nutritional strategies, using a mathematical model. This model incorporates four nutritional types commonly found in planktonic ecosystems: (1) heterotrophic bacteria that consume dissolved organic matter and are prey to some of the other organisms; (2) heterotrophic zooflagellates that depend entirely on bacterial prey; (3) phototrophic algae that depend only on light and inorganic nutrients, and (4) mixotrophs that photosynthesize, take up inorganic nutrients, and consume bacterial prey. Mixotrophs are characterized by a parameter representing proportional mixing of phototrophic and heterotrophic nutritional strategies. Varying this parameter, a range of mixotrophic strategies was examined in hypothetical habitats differing in supplies of light, dissolved organic carbon, and dissolved inorganic phosphorous. Mixotrophs expressing a wide range of mixotrophic strategies persisted in model habitats with low phosphorus supply, but only those with a strategy that is mostly autotrophic persisted with high nutrient supply, and then only when light supply was also high. Organisms representing all four nutritional strategies were predicted to coexist in habitats with high phosphorus and light supplies. Coexistence involves predation by zooflagellates and mixotrophs balancing the high competitive ability of bacteria for phosphorus, the partitioning of partially overlapping resources between all populations, and possibly nonequlibrium dynamics. In most habitats, the strategy predicted to maximize the abundance of mixotrophs is to be mostly photosynthetic and supplement nutritional needs by consuming bacteria.

A mechanistic explanation for pH-dependent ambient aquatic toxicity of Prymnesium parvum carter.

The harmful algal bloom species Prymnesium parvum has caused millions of dollars in damage to fisheries around the world. These fish kills have been attributed to P. parvum releasing a mixture of toxins in the water. The characterized toxins, reported as prymnesin-1 and -2, have structural similarities consistent with other known ionizable compounds (e.g., ammonia). We investigated whether pH affects the toxicity of P. parvum under conditions representative of inland Texas reservoirs experiencing ambient toxicity from bloom formation. We evaluated pH influences on toxicity in laboratory and field samples, and modeled the physicochemical properties of prymnesins. Aquatic toxicity to a model fish and cladoceran was reduced by lowering pH in samples obtained from reservoirs experiencing P. parvum blooms; similar observations were confirmed for experiments with laboratory cultures. A pKa value of 8.9 was predicted for the prymnesins, which suggests that ionization states of these toxins may change appreciably over surface water pH of inland waters. These findings indicate that ionization states of toxins released by P. parvum may strongly influence site-specific toxicity and subsequent impacts to fisheries. Consequently, these results emphasize the importance of understanding processes that affect pH during P. parvum blooms, which may improve predictions of ambient toxicity.

Competition and coexistence in flowing habitats with a hydraulic storage zone.

This paper examines a model of a flowing water habitat with a hydraulic storage zone in which no flow occurs. In this habitat, one or two microbial populations grow while consuming a single nutrient resource. Conditions for persistence of one population and coexistence of two competing populations are derived from eigenvalue problems, the theory of bifurcation and the theory of monotone dynamical systems. A single population persists if it can invade the trivial steady state of an empty habitat. Under some conditions, persistence occurs in the presence of a hydraulic storage zone when it would not in an otherwise equivalent flowing habitat without such a zone. Coexistence of two competing species occurs if each can invade the semi-trivial steady state established by the other species. Numerical work shows that both coexistence and enhanced persistence due to a storage zone occur for biologically reasonable parameters.