Introduction

At Nexa Labs, we are developing the world’s most powerful cattle monitoring solution. Our system is driven by in-ear subcutaneous microchips that track deeper animal biometrics than anything else on the market. Although similar devices have been prototyped for nearly a decade, they have not been successfully harnessed for the benefit of real-world cattle operations. This has been due to the lack of “productization,” ie: addressing implementation nuances that allow subcutaneous microchips to be used as part of a complete and long-lasting product that delivers actionable insights to downstream users. Nexa is closing this gap with an end-to-end monitoring solution that finally realizes the potential of in-ear microchips. This overview reviews literature that scientifically rationalizes the placement and capability of our microchip, as well as explore the advantages of our strategy versus existing wearables (ex: collars, ear-tags) and bolus devices.


Biometric Capabilities of Nexa Microchips vs Collars, Ear-Tags, and Boluses

The accuracy of cattle insights (ex: heat identification and early disease detection) is determined by the quantity and quality of biometrics analyzed by monitoring solutions. Reasonably, algorithms that predict disease or heat status from biometric data perform better if they have access to more high quality inputs. Thus, it is important to consider the types of data that different monitoring solutions can capture.


Activity and Behavior Monitoring

The vast majority of monitoring devices utilize an accelerometer that tracks animal movement patterns. Accelerometer movement data can then be processed by different algorithms to generate features related to animal activity and behavior (ex: eating time, resting time, rumination time, etc). For instance, low-frequency patterns may reflect rumination activity, while high frequency movements may indicate sudden shock or excitement. Such features underpin key symptoms of events highly relevant for cattle management; for example, cows tend to exhibit increased activity during estrus [1]. Lee & Seo [2] comprehensively reviewed literature on the performance of external wearables and bolus devices for inferring activity and behavior. Across eating, rumination, water-drinking, standing, walking, and step count, these devices demonstrated strong correlation with ground-truth measurements recorded through visual observation or gold-standard testing protocols. The primary exception was drinking time, reflecting a weak distinct association between drinking behavior and cattle head or neck movement. Other studies, such as from Rahman et al [3], have also found comparable activity and behavior analysis capability across monitoring devices. In the Nexa system, activity and behavior features are derived from the ear-tag module paired with the implanted microchip for power and communication. Thus, as a baseline, the performance of the Nexa system on these tasks is at minimum on-par with that of existing ear-tags, independent of algorithmic improvements. However, note that Nexa has already achieved superior model performance on benchmark datasets due to improved algorithmic design.


Temperature Monitoring

Body temperature demonstrates strong predictive capability for both heat detection [4,5] and health issue identification [6–8]. Rectal temperature is typically used as the gold standard measurement, though other measurement sites have also been investigated due to the practical difficulty of continuous rectal monitoring. Generally, rumen bolus temperatures track rectal measurements closely, and follow similar trends under cattle conditions with relatively minimal interference from ambient environmental effects [9–11]. Collar devices are unable to capture animal temperature due to their loose adherence to the cattle skin. Many ear-tag based monitoring systems claim to track temperature through sensors placed into contact with the exterior surface of the ear. These largely rely on tracking changes in measured temperature over time and comparing individuals to the herd average. However, external surface temperatures are highly prone to fluctuation driven by ambient environmental conditions such as temperature and humidity [12]. Among surface-level temperature monitoring, the ear is less robust compared to other sights which are more difficult to monitor using existing wearable technology such as the forehead and eyes. This means that downstream insights derived using temperature measured by wearables may be prone to error under non-optimal conditions such as heightened environmental temperature or irregular weather patterns. Furthermore, reliance on comparison to herd averages may cause these solutions to fail when herd-wide conditions change rapidly, such as in the case of disease outbreak. Tracking temperature through subcutaneous in-ear sensors – as the Nexa system does – provides more robust measurements compared to external wearables. This is supported by literature that demonstrates continued accuracy of subcutaneous temperature readings under diverse conditions, as well as the potential for patterns in these readings to enable early disease and heat identification [13–16]. These advantages compared to external wearables are enabled by the insulation granted by thick cattle skin which provides a natural barrier between internal physiological conditions and the ambient environment.


Physiological Signals Beyond Temperature

Given that cattle physiology is intertwined with health and reproductive conditions, it is highly favorable to monitor further biometrics beyond solely activity and temperature (ex: heart rate, respiration rate, blood oxygen, blood pressure, etc) [17–19]. Some bolus devices track pH condition, but this leads to sharp decreases in battery life that severely limit scalable applicability in commercial dairy settings [20]. Existing commercially available external wearables are incapable of tracking biometrics beyond activity and temperature from outside animal skin. Though some speculate it may one day be possible to monitor an expanded physiological parameter-set from an ear-tag design, such aspirations encounter significant fundamental roadblocks such as fur thickness interfering with optical sensors [21]. The Nexa subcutaneous in-ear placement however, lends itself to monitoring diverse physiological signals due to the vascularization of the cattle ear including large veins, arteries, and capillaries. Similar regions have successfully been monitored for such biometric tracking in other organisms including humans, rats, pigs, and more [22–24].


Utility and Practicality of Cattle Monitoring Solutions

The purpose of monitoring cattle biometrics is to provide actionable information to key farm stakeholders including operation owners, managers, veterinarians, and consultants. The primary items of endpoints are early disease detection – at a stage when either treatment or preventative measures can be readily applied – and identification of animal heat or estrus. Thus, robustness on these tasks is key to judging the utility of monitoring solutions. Due to the previously discussed limitations surrounding scale and quality of biometric monitoring, wearable devices are plagued by degraded performance under non-optimal conditions. For example, environmental factors like temperature, humidity, and weather phenomena can lead to false positives in disease and heat detection [25]. Weinert-Nelson et al [26] recently demonstrated this issue with the CowManager ear-tag system; in their study, the authors found significant decreases in the accuracy of CowManager-derived insights under standard heat stress conditions. Other studies have found similar results for collars and various ear-tag systems - in many cases, there are no standardized validation efforts published or conducted commercially available systems, potentially due to negative perception surrounding subpar performance [27]. Studies that do claim to validate these systems often fail to report relevant metadata such as weather conditions and animal management practices [28]. Rumen bolus devices have demonstrated more accurate and consistent capability for disease and heat detection, compared to wearables like collars and ear-tags [6,29,30]. This is largely due to their ability to measure internal body temperature with greater fidelity. The Nexa system is expected to mirror and improve upon these strengths, as it too assesses body temperature from inside the animal, rather than externally. Furthermore, the Nexa microchip measures physiological signals on-par or beyond those that bolus devices can track.


Practicality of Solution Implementation on Dairy Operations

Though their performance is markedly worse for key tasks, wearables including collars and ear-tags have thus far experienced stronger adoption than bolus devices [31]. This is due to practical application considerations where bolus devices lag behind. For instance, as aforementioned the battery life of boluses is severely as far as less than a year when they measured pH in addition to activity and temperature. Their high production cost also results in downstream pricing that is prohibitively expensive for medium to large-sized commercial dairy operations. This is particularly difficult to rationalize for implementation when considering the non-rechargable and non-retrievable nature of boluses; when they run out of battery, a new device must be inserted for continued function. Finally, as bolus devices are placed in the main body of animals, cattle with these devices inserted are typically not suitable for meat processing (ex: at end-of-life) due to concerns with electronic devices being run through processing equipment. This reduces the applicability of boluses for beef cattle and dairy cows which are terminally processed. Unlike bolus devices, the in-ear placement of the Nexa microchip allows implanted cattle to follow standard meat processing procedures, as the ear is typically removed as one of the first steps. Furthermore, all power requirements for the microchip are met through interfacing with the paired ear-tag, alleviating the need to insert new devices in the lifetime of the animal. The miniature form-factor also enables low production costs that translate to pricing aligned with external wearables. Thus, the Nexa system is as widely applicable as wearables, with increased performance on-par with or outperforming bolus devices.


Conclusion

Overall, this review explored the advantages and drawbacks associated with commercially available cattle monitoring solutions. Our findings demonstrated that the design of the Nexa system positions it for a maximized combination of capability, accuracy, and practical applicability. Further analyses will investigate field-validation data from implementation of the Nexa system on live animals.