The program has been quite successful; peak molinate concentrations have been reduced by better than 90 percent in the Sacramento River and Colusa Basin Drain since the beginning of monitoring and thiobencarb concentrations have declined as well, although by a lesser percentage . It should be noted that peak concentrations are an imperfect indicator as weather events and other sources of variability can significantly affect detectable levels of these chemicals. The environmental and health effects of leaching and runoff of chemicals in U.S. agriculture are difficult to quantify, primarily due to measurement issues and uncertainty. Approaches to estimation of these external costs vary and include using abatement or clean-up costs directly; developing proxy variables for environmental damage, including pounds of AI applied or indices of chemical properties; and assuming a dollar value per unit of damage. Contingent valuation methods have also been used to estimate consumers’ willingness-to-pay to avoid exposure through risk reduction . Although admittedly simplistic, one proxy for environmental damage is total pounds of chemicals applied. Based on 2002 acreage and use figures from DPR and the chemical labels, large plastic plants for pots approximately 17.2 million pounds of herbicides were applied that year with 3.86 million pounds of active chemical ingredient.

As shown in Table 3, total pounds of chemical herbicides applied per acre are expected to decrease by at least 84 percent with adoption of the HT system and total poundage of AI is predicted to decrease by at least 87 percent under the two-treatment scenario. Cultivation of HT rice could thus decrease total herbicide poundage by between 7.27 and 10.9 million pounds and AI poundage by between 1.69 and 2.53 million pounds, assuming 50 to 75 percent adoption. However, this simple measure ignores toxicity, mobility, and persistence of different chemicals in the soil and water that are likely to significantly affect external damage costs . Similarly, the mandatory water-holding periods currently in place are designed to dissipate the damage done by conventional chemicals. While this study makes no attempt to further quantify the reduced chemical damages from adoption of transgenic rice, several other projects have addressed the relationship between water quality and HT crops and are summarized in Gustafson. Computer-simulation models used by the U.S. Environmental Protection Agency have predicted lower levels of chemical concentrations in runoff from transgenic corn systems than from conventional corn production. Furthermore, the herbicides used in the transgenic system have a “favorable chemical profile” in that EPA’s water-quality standards allow for greater concentrations of these chemicals in water than the traditional herbicides used in conventional corn production . Case studies cited in Gustafson for Bt cotton, HT corn, and HT soybeans confirm these results since concentrations of herbicides in watersheds were well below standards for a number of diverse geographic areas. We conclude that, while most, if not all, pesticides applied in agricultural systems introduce some degree of risk and thus potential damage to the environment, the reduced application rates and chemical properties of glyphosate and glufosinate have the potential to further reduce external damage costs from rice production.

In addition, production benefits, specifically in terms of yields, may be enhanced if the holding period for floodwater is reduced for the transgenic system due to the lowered toxicity of the associated chemicals. This is in accordance with previous studies that concluded that cultivation of transgenic crops, in general, is consistent with increased environmental stewardship . However, it should be noted that these conclusions are based on the assumption that weeds resistant to currently available chemical controls in California do not exhibit this property towards glyphosate and glufosinate. As with many chemical agents, repeated applications of the same AI on the same plot may result in self-selection of weed varieties resistant to that ingredient, thus potentially reducing the environmental benefits of transgenic-crop cultivation in the long run as producers increase applications or shift to alternative means of control.This study has used a static, partial-budgeting approach to estimate the potential net economic grower benefits associated with adoption of one cultivar of GM rice in California. Scenarios were developed based on average cost data and actual pesticide-use data, as well as on a three-year field study of herbicide resistant weeds. Sensitivity analysis was conducted using both deterministic and stochastic methods to represent heterogeneity across growers and uncertainty regarding modeling assumptions. The results suggest that a production strategy including GM rice varieties could lead to significant economic benefits for many growers in at least the near term.

Those most likely to benefit from adoption of transgenic rice are growers with relatively high herbicide material and application costs, likely as a result of weed resistance, and those who are restricted to certain chemical agents as a result of state or national regulations. Field-trial results suggested that a transgenic weed-management strategy over multiple years is competitive with a rotational strategy under certain assumptions and dominates a continuous-molinate and intensive-herbicide regime. These findings are generally consistent with ex post transgenic crop analyses for corn, soybeans, and cotton, most of which show positive or neutral economic benefits from adoption . In addition, water quality degradation is not likely to occur with transgenic rice adoption as chemical-application rates are expected to sharply decline and the toxicity of the associated chemicals is generally less than more traditional herbicides. We must point out that this study has certain limitations. First, these results are based in large part on ex ante assumptions regarding outputs and inputs, especially the relatively lower cost of glufosinate herbicide per gallon relative to the alternatives. It is expected that Bayer CropScience will set the price of glufosinate in accordance with its portfolio of transgenic crops , so the introduction of transgenic rice will not affect the price of this herbicide. However, this is far from certain. Additionally, this study does not account for the response of other chemical producers who may change their pricing strategy in response to the introduction of HT rice cultivars. Second, this analysis is static or, in the case of the field trial, a sequence of static analyses. Thus, several important elements that have been purposely excluded may impact the conclusions. Among these is the lack of dynamic effects in the model, such as the potential of glufosinate resistance in rice fields through the natural selection process or the effects on watergrass or other weed seedbanks . There has been little evidence of such resistance in the literature; nevertheless, it may become more of an issue if significant adoption occurs. Third, we assume that GM rice will be accepted by the marketplace, that only a small share of the market will be willing to pay a non-GM price premium, and that the costs of segregating non-GM from GM rice will be modest. By the time GM rice is adopted in California, the political opposition to biotechnology will have likely declined in the state and elsewhere. More importantly, California may not be the first growing region in the world to adopt transgenic rice. If GM rice is first adopted in Asia, then adoption in California will have only a small impact on the market. Clearly, more research is required on the market response to GM rice. Finally, general-equilibrium price and quantity effects that impact both consumer and producer welfare are not included in this analysis. In a highly differentiated market like the rice market in California, such effects are most likely at the aggregate market level and would impact the estimation of the total surplus lost from banning GM rice. However, we chose to focus our analysis at the agent level via the partial-budgeting approach and we reported sensitivity results assuming exogenous market conditions. Therefore, our model does not require forecasting of potential adoption rates or estimating systems of regional supply and demand for the conventional and biotech markets, plant pots with drainage thus considerably simplifying the analysis.Traumatic Brain Injury is a form of acquired brain injury caused by external impact to the head that results in damage to the brain. It is a common cause of death and disability in the United States and can be caused by a variety of factors including falls, motor vehicle crashes, sports, or combat injuries.

TBI affects an estimated 2 million people in the U.S. across all age groups, according to data from U.S. TBI often leads to neurological problems in individuals, including cognitive, motor, and sleep-wake dysfunction. Currently available medical tools for TBI diagnosis are largely subjective and a lack of consensus regarding what constitutes mild TBI adds to the complication of the under-diagnosis of mTBI. TBI is categorized into mild, moderate, or severe based on the Glasgow coma scale , Loss of consciousness , and Post-traumatic amnesia which are qualitative tests rather than quantitative measures. The World Health Organization’s definition of mTBI allows for a GCS score of 13–15 to be assessed after the typical 30-min timeframe, which accounts for the expected time of arrival of a qualified healthcare provider. However, the GCS has its drawbacks. Being highly inter observer dependent makes it necessary to report exact findings rather than just the score. In addition, one of the key parameters in GCS is the eye score which might be unattainable in case of an eye injury. Considering the requirements from a medical resource standpoint, existing clinical tools used to diagnose mTBI such as Magnetic Resonance Imaging and Computer Tomography require an extensive, high-cost clinical setup and specialized operator skill set which are not always available at the time and place of an incident, and these neuroimaging tests may still be negative in many cases of mTBI. As a result of the limitations of the present-day methods used to detect TBI, there is a need for new technology capable of rapid, accurate, noninvasive, and most importantly, field-capable detection of mTBI to bridge the technological gap that exists today. Early, objective, and reliable mTBI detection can help affected individuals undergo timely monitoring and therapy and can prevent death in severe cases. Machine learning techniques provide a way to study mTBI and create systems to help objectively diagnose and monitor mTBI presence and stages in individuals. Recently, machine learning techniques have been investigated for the purpose of classifying mTBI from electroencephalogram data in mice based on models created using the lateral Fluid Percussion Injury method. FPI induced mice demonstrate very similar behavioral deficits and pathology to those found in humans afflicted with mTBI, including sleep disturbances. In this work, we use EEG data acquired from the compelling FPI mouse model of mTBI. This work focuses on creating a machine learning-based, fast, portable, and ready to use EEG classification system for mTBI detection using Raspberry Pi 4 . This system works with a variety of machine learning models and can be used along with live EEG recording systems to detect mTBI. This capability can enable field use and make early mTBI detection possible without the requirement of extensive medical setup or specialized medical domain knowledge. Further, this work has the potential to create avenues for implementing mTBI related real-time connected health monitoring systems and allow further research on real-time mTBI detection using EEG. The deployment system created in this work incorporates an Analog to Digital Converter front-end and utilizes Convolutional Neural Network and XGBoost predictive models to perform sleep staging and detect the presence of mTBI using a single-channel EEG signal. We used spectral features of the EEG signal for our models which were shown to be suitable for detecting mTBI in mice with an accuracy of more than 80%. The system captures and classifies EEG epochs into four target classes—Sham Wake, Sham Sleep, mTBI Wake, and mTBI Sleep. We demonstrate that our system can capture physical EEG signals and perform feature extraction and prediction using the XGBoost model in the order of 0.02 s per epoch which makes it possible to quickly detect the presence of mTBI. We also verify that the cross-validation metrics obtained on the RPi based system are identical to those obtained on a High-Performance Computer such as a 64-bit workstation computer running macOS or Windows. The Related Works subsection covers previous relevant investigations in this area. We describe the deployment system design, operation, classification model configuration, and validation techniques in the Methods section.