Electrochemical measurements play a crucial role across various domains including air quality assessment, biological analysis, and the food industry. Miniaturized and power-efficient electrochemical potentiostats, facilitated by integrated circuits, have been instrumental in enabling wearable devices. However, the traditional CMOS potentiostat designs limit the simultaneous measurement of multi-sensor arrays within the same electrolyte solution where different bias potentials need to be applied to each sensor. Moreover, the utilization of modern integrated circuits with low supply voltage limits the applicability of electrochemical reactions requiring higher potential windows. This paper introduces an innovative potentiostat architecture that allows for sharing the reference electrode among multiple channels as well as expanding the voltage range for electrochemical cells. The novel potentiostat design is thoroughly described and results from a PCB prototype show strong output current linearity. Furthermore, verification of simultaneous measurements at different electrochemical bias conditions is presented and shown to match results from a commercial instrument.

Face-to-face interactions provide an essential ingredient to the development of relationships that are critical to human well-being. Nonverbal cues, such as gaze direction and body language, play a large role in interpersonal exchanges, but they are poorly understood. To provide new insight into nonverbal cues, this paper presents a compact, wearable, real time social behavior monitoring system and details a methodology for extracting a key nonverbal cue called facing time, the amount of time one person faces another during social interactions. The design of an IMU-based embedded device, called Earloop, is discussed along with package design for the ear-mounted wearable. The algorithm developed for real time assessment of facing time is presented, including a unique method employing hysteresis to deal with cases where facing direction windows of different users are overlapping. Test results are presented that demonstrate viability of an IMU-based system for monitoring facing time. Experimental evaluations under balanced, unequal, and disengagement conditions yielded an overall average error of 1.79%, demonstrating high accuracy in real-time monitoring. These results support the concept of using a head-worn wearable to monitor social interactions without the use of cameras and demanding signal processing algorithms. By expanding this system to monitor additional nonverbal cues, this paper lays the groundwork for a future wearable capable of assessing a wide range of social behaviors in real time.

Modern sensor technologies have been employed to monitor aspects of social interactions, such as human emotions, that are known to influence human health. To reduce the negative impact that some behaviors could have on our health, real-time monitoring of social interactions is desired to bring awareness of human behavior through real-time feedback. Still, the design of systems for the real-time monitoring of social interactions poses considerable challenges that range from multi-sensor integration to signal analysis. Intending to overcome these challenges, this paper presents a study of a variety of sensor modalities and the design of a multi-sensor framework that allows the study and real-time analysis of both in-person and virtual social interaction environments. The framework consists of three multi-sensor nodes and a central unit. Results show validation of the variety of sensor data collected from a single sensor node and behavioral information that can be identified due to data synchronization from multiple sensor nodes.

The field of human augmentation (HA) has received an increased research focus from neurotechnologists due to advancements in information sensing and processing capabilities. Despite advancements in HA, machine-to-human communication (MHC) is still reliant on visual and auditory pathways. Electrotactile stimulation of peripheral nervous system can be used as an alternative pathway for MHC, contingent upon consistently distinct electrotactile sensations can be created. A person can be trained to associate a distinct electrotactile sensation with a specific information cue, effectively creating an alternative pathway for MHC. A technique for creating distinct electrotactile stimulation is modulating the intensity of the observed electrotactile waveform. Among the possible techniques to modify intensity, this paper investigates the possibility of manipulating a train of high-frequency pulses with the duty cycle of a superimposed mask. Furthermore, this paper explores the limits of human ability to distinguish intensity levels created through manipulation of high-frequency pulses. This paper found that using the presented frequency manipulation technique makes it possible to reliably create 3 intensity levels with an accuracy higher than 85%.

MOTIVATION: Next-generation sequencing has become exceedingly common and has transformed our ability to explore nonmodel systems. In particular, transcriptomics has facilitated the study of venom and evolution of toxins in venomous lineages; however, many challenges remain. Primarily, annotation of toxins in the transcriptome is a laborious and time-consuming task. Current annotation software often fails to predict the correct coding sequence and overestimates the number of toxins present in the transcriptome. Here, we present ToxCodAn, a python script designed to perform precise annotation of snake venom gland transcriptomes. We test ToxCodAn with a set of previously curated transcriptomes and compare the results to other annotators. In addition, we provide a guide for venom gland transcriptomics to facilitate future research and use Bothrops alternatus as a case study for ToxCodAn and our guide. RESULTS: Our analysis reveals that ToxCodAn provides precise annotation of toxins present in the transcriptome of venom glands of snakes. Comparison with other annotators demonstrates that ToxCodAn has better performance with regard to run time ($>20x$ faster), coding sequence prediction ($>3x$ more accurate) and the number of toxins predicted (generating $>4x$ less false positives). In this sense, ToxCodAn is a valuable resource for toxin annotation. The ToxCodAn framework can be expanded in the future to work with other venomous lineages and detect novel toxins.

Monitoring human emotions through wearable systems has become an important area of research. Electrodermal activity (EDA) has proven to be a good indicator of emotional arousal, and numerous works have focused on using EDA data to predict emotional states. However, to successfully integrate EDA data into real-time wearable emotion recognition systems, several challenges of practical real-life scenarios, need to be addressed. This paper explores the relationship between speech signals and EDA reactions and analyzes a new approach for classification of skin conductance reactions elicited by emotional arousal using speech signals as a triggering event. Results show an average improvement in skin conductance reaction classification accuracy of at least 5.6% when using speech-triggered reactions compared with traditional methods. The use of speech as a triggering event could help improve real-time emotion recognition algorithms implemented within wearable systems.

Wearable real-time systems such as health monitors exhibit a need for user feedback capable of communicating a wide range of messages. This paper describes how high-frequency (100 - 2k Hz) electrotactile stimulation pulses can be utilized for producing multiple unique message sensations within a limited time period of 0.5s. Two experiments were conducted on willing participants using a custom electrotactile stimulator. Experiment 1 investigated the effectiveness of producing unique sensations by varying electrotactile pulse frequency above 100Hz. Results indicate that even though pulse frequencies above 100Hz produce detectable sensations, discrimination by frequency higher than 100Hz is not feasible. Experiment 2 investigated the effectiveness of discriminating electrotactile pulse frequencies higher than 100Hz when the signal was modulated at a fixed low frequency (6Hz) “bundle”. Results show that variations in low-frequency Bundle duty-cycle produced at least three distinct sensations that may be utilized to expand the available set of uniquely perceived electrotactile sensations.

Deterministic lateral displacement (DLD) devices have demonstrated great promise in separation of micro and nano-sized particles, with important applications in biomedical research and healthcare monitoring. This paper introduces a new cascaded multi-section DLD approach toward expanding the dynamic range of particle sizes separated. A robust model has been developed to analyze the design tradeoffs and practical fabrication limits of this new approach. Results show that by cascading multiple sections of increasingly smaller gap size and critical separation dimension, a wide spectrum of size fractionation dynamic ranges and minimum separation resolutions can be achieved. Moreover, the presented model allows designers to visualize the cost of achieving various performance goals in terms of overall device size. Model results based on DLD theoretical equations are first presented, followed by model results for both circle and I-shaped pillar options that apply scaling restrictions associated with their practical fabrication limits.

Abstract Computing the frustration index of a signed graph is a key step toward solving problems in many fields including social networks, political science, physics, chemistry, and biology. The frustration index determines the distance of a network from a state of total structural balance. Although the definition of the frustration index goes back to the 1950s, its exact algorithmic computation, which is closely related to classic NP‐hard graph problems, has only become a focus in recent years. We develop three new binary linear programming models to compute the frustration index exactly and efficiently as the solution to a global optimization problem. Solving the models with prioritized branching and valid inequalities in Gurobi, we can compute the frustration index of real signed networks with over 15 000 edges in less than a minute on inexpensive hardware. We provide extensive performance analysis for both random and real signed networks and show that our models outperform all existing approaches by large factors. Based on resolution time, algorithm output, and effective branching factor we highlight the superiority of our models to both exact and heuristic methods in the literature.

Abstract This article reports on an approach for the VeRoLog Solver Challenge 2019: the fourth solver challenge facilitated by VeRoLog, the EURO Working Group on Vehicle Routing and Logistics Optimization. The authors were awarded third place in this challenge. The routing challenge involved solving two interlinked vehicle routing problems for equipment: one for distribution (using trucks) and one for installation (using technicians). We describe our solution method, based on a matheuristics approach in which the overall problem is heuristically decomposed into components that can then be solved by formulating them as set partitioning problems. To solve these set partitioning problems we introduce a novel method we call “columnwise neighborhood search,” which allows us to explore a large neighborhood of the current solution in an exact manner. By iteratively applying mixed‐integer programming methods, we obtain good quality solutions to our subproblems. We then use a simple local search “fusion” heuristic to further improve the solution to the overall problem. Besides introducing and discussing this solution method, we highlight the problem instances for which our approach was particularly successful in order to obtain general insights about our methodology.

High channel count neural implants that can record brain activities across diverse cortical regions represent the next step toward whole brain interfaces that will enable new understanding of brain operation and treatment of many neural disorders. To overcome the size and power constraints limiting the channel count of existing neural implants, this paper presents a new neural amplifier array design that utilizes hardware sharing to achieve low power and compact size. Moreover, to ease the burden of large volume data handling, in-situ data decimation is performed to enable off body evaluation of synchrony between signal pairs. A 32-channel analog front end array was designed and post-layout simulations show that the entire front end occupies only 0.031 mm2per channel while consuming only 3.34 μW per channel at 3.3 V in 0.5 μm CMOS. This front end decimates data by an order of magnitude while keeping the synchrony information with more than 89.1% accuracy.

This work presents a computationally efficient real-time adaptive clustering algorithm that recognizes and adapts to dynamic changes observed in neural recordings. The algorithm consists of an off-line training phase that determines initial cluster positions and an on-line operation phase that continuously tracks drifts in clusters and periodically verifies acute changes in cluster composition. Analysis of chronic recordings from non-human primates shows that adaptive clustering achieves an improvement of 14% in classification accuracy and demonstrates an ability to recognize acute changes with 78% accuracy, with significantly improved computational efficiency compared to the state-of-the-art. The presented algorithm is suitable for long-term chronic monitoring of neural activity in many applications of neuroscience research and control of neural prosthetics and assistive devices.

A highly flexible software platform for automated data acquisition, production of research objects with data provenance and curation of experiment results throughout the life cycle of the data will be introduced and demonstrated along with a custom miniaturized electrochemical sensor system. The software platform, called eGor, allows users to define test procedures and components through a user friendly access point. Specific experiment definitions can be saved for later use or recalled as templates for modified tests. The user can then remotely execute the test on any connected physical experiment workbench in a real lab environment. Once the test is complete, eGor will capture an organized and metadata-rich research object that includes the raw test data, detailed definition of the test setup, and all procedural elements of the executed test such as the timings, test successions, device conditions, etc. The generated research object are stored for subsequent curating or data analysis, and any access or treatment of the results is automatically recorded to maintain data provenance. eGor also allows stored research objects to be shared with collaborators or provided to any institution that would be interested in reproducing the same results. The results may be inspected at various levels of detail, annotated and compared with other research objects. Thus, eGor would help researchers increase their confidence in their results and conclusions and promote improved research reproducibility. During this demonstrations the users will be allowed to interact with hardware blocks that mimic a simple lab setup, and the users may define, schedule and perform tests using these devices and save their test results as research objects.

Airborne particulate matter (PM) pollution, especially fine particles with a diameter of 2.5 μm or smaller (PM2.5), has caused severe air quality issues that threaten human life and contribute to global mortality. Thus a low cost, portable or wearable platform for individual PM2.5 monitoring is of great interest. This paper introduces a platform for portable real-time PM2.5 monitoring that combines a particle separation microfluidic channel and electrochemical detection. The microfluidic device utilizes an I-shape pillar based deterministic lateral displacement method to achieve high separation efficiency of different particle sizes. Electrochemical detection was implemented for particles measurement to achieve high sensitivity and simplify the instrumentation compared to conventional optical methods. Both separation and detection results show that this platform is a promising option for portable PM2.5 monitoring.

In depth characterization of nanopores such as ion channel proteins holds great value for medical and pharmaceutical applications. In this paper, an electrochemical interface circuit (EIC) is presented that enables both readout of individual nanopores and high throughput implementation within a lab-on-CMOS array platform. The EIC was designed for an electrochemical array microsystem that would facilitate proteomics research via parallel characterization of multiple ion channel proteins with single channel resolution. Fabricated in 0.5 μm AMI CMOS, the EIC can record rapid and extremely weak ion channel current pulses within a very noisy environment. Measurements show that the circuit can detect currents as low as 10 pA with a pulse width of 10 μsec. The readout circuit is low power and very compact to facilitate up to 1024 channels on a large CMOS chip.

The integration of biosensors, microfluidics and CMOS instrumentation provides a compact lab-on-CMOS microsystem well suited for high throughput measurement. This paper describes a screen-printed planar metallization technique for lab-on-CMOS that overcomes challenges associated with traditional thin film metallization. Utilizing a chip-in-carrier packaging approach with an epoxy carrier, screen-printed electrical interconnects are shown to reliably resolve up to 10μm step height differences between the CMOS chip and the surrounding carrier that supports microfluidics. The metallization process presented in this paper is also shown to be compatible with subsequent microfluidic integration to complete the lab-on-CMOS device platform.

Chronoamperometric gas sensor arrays show great promise for ultra-low power consumption and low cost for wearable gas sensing devices for human safety and health monitoring. This paper presents a novel power efficient instrumentation circuit with high sensitivity and large dynamic range for wearable chronoamperometric gas sensor arrays. This instrumentation combines an input digital modulation technique and a semi-synchronous incremental ΣΔ ADC structure to achieve very high power efficiency over a large dynamic range with high sensitivity. The proposed instrumentation was implemented in 0.5 μm CMOS technology. Measurement results demonstrate that 164dB cross-scale dynamic range and 100 fA sensitivity are achieved with a high power efficiency.

Because membrane proteins are critically important in biological processes and among the most prevalent drug targets, a thorough understanding of their structure and function is highly desired. However, this has been difficult to achieve due to the laborious nature of existing tools and techniques for membrane protein studies. This paper presents a new microsystem concept for high throughput membrane protein characterization utilizing an array of synthetic planar lipid bilayers within multi-channel microfluidics on the surface of CMOS electrochemical instrumentation circuits. A new technique for CMOS polymer packaging that enables planar routing of bonding pad signals and formation of multi-channel microfluidics on the surface of on-chip electrochemical sensor electrodes is demonstrated. The developed techniques enable ~1000 planar lipid bilayer membrane protein interfaces to be implemented directly on the surface of high performance CMOS instrumentation and the integration methods can be used in lab-on-chip applications.

Future brain machine interface systems will require recording thousands of neural channels, making it important to minimize the power and area of neural interface integrated circuits. Spike detection is an essential step for neural signal processing. This paper describes the design of a spike detection circuit based on the nonlinear energy operator (NEO) algorithm that is optimized for power and area. Through statistical analysis of NEO coefficients, the number of computations is minimized and the number of registers is shown to be as low as one per channel without degenerating spike detection performance. Based on an analysis of the power-area tradeoff, an optimal 16-channel interleaved architecture is presented and shown to achieve a factor of 4 improvement in power-area product compared to reported NEO implementations.

Integrated sensor arrays on CMOS instrumentation chips are attractive to many biological and biomedical sensor applications. However, the packaging of CMOS circuitry for use within a liquid environment remains as an open challenge. Reliable post-CMOS electrode fabrication and packaging processes that are critical to the development of integrated electrochemical biosensors are presented in this paper. A die-level photolithography process was developed that provides uniform coverage for accurate patterning over 87% of a 3×3mm2silicon substrate. In addition, a new process has been developed to package post-CMOS fabricated electrode arrays. This etchless parylene packaging reduces processing time and improves fabrication yield. These techniques enable realization of on-CMOS biosensors operating in liquids.

A monolithic chemiresistor (CR) array micro-system with microfluidic channel for micro gas chromatograph (μGC) is presented in this paper. A CMOS readout chip was designed for amplifying and conditioning the signal of a 4×2 MPN-coated CR array fabricated on the surface of the CMOS chip. A micro glass lid with input and output capillary tubes was developed as a gas flow channel and mounted on the CMOS array, providing an interface to a micro flow column in a μGC platform. After all of the CMOS-compatible processing, the monolithic CR array was tested within a GC platform, and both the CR array and circuit were demonstrated to function as designed.

Inductively coupled transceivers (ICT) are widely being employed in implantable biomedical devices for wireless communication with the external world. The performance of such implant-to-body-surface communication is dependent on the characteristics of the transmitter, the receiver and the channel between them, which consists of layers of biological tissue. In this paper a representative ICT has been studied to characterize the effect of different physical orientations on the bit error process. The effect of biological tissue on the communication channel has also been quantified. The channel was measured to have a bit-level memory of 3 bits and a packet level memory of 6 packets. The impact of forward error correction (FEC) was analyzed and the use of 3-bit FEC was found to reduce the number of retransmissions by 65%. The results of this study enable optimization of reliable implantable communication systems.

Nanoparticle coated chemiresistor (CR) arrays enable highly sensitive vapor detection in systems such as a micro gas chromatograph or an electronic nose. However, they suffer shortcomings such as a small response compared to a large baseline value, large baseline variation across devices, and significant baseline drift over time. This paper describes a new high-resolution CR array readout circuit with adaptive baseline control. The 8-channel readout circuit occupies 2.2mm × 2.2mm in 0.5μm CMOS technology, consuming 66μW per channel from a 3.3V power supply. It achieves a worst-case resolution of 125ppm over a baseline resistance of 60kΩ to 10MΩ, equivalent to 120dB dynamic range.

Multi-channel neural signal recordings need high data compression and efficient data transmission. Our previous work has shown a practical data compression solution based on discrete wavelet transform, multi-level thresholding and run length encoding. This paper presents a custom designed communication protocol for bidirectional data telemetry to and from the implanted module. A global controller is also presented which configures, operates and unites all the modules together effectively and efficiently into a 32-channel system. Performance of the communication protocol and the compression engine is analyzed.

Impedance spectroscopy (IS) is a powerful tool for characterizing materials that exhibit a frequency dependent behavior to an applied electric field. This paper introduces a fully integrated multi-channel impedance extraction circuit that can both generate AC stimulus signals over a broad frequency range and also measure and digitize the real and imaginary components of the impedance response. The circuit was fabricated in 0.5μm CMOS and consumes 355μW at 3.3V. Tailored for protein and lipid bilayer characterization, the signal generator produces sinusoidal waves from 1Hz to 10kHz. To suit a variety of applications, the impedance extraction circuit provides a programmable current measurement range from 100pA to 100nA with a measured resolution of ~100fA. Occupying only 0.045mm2per measurement channel, the circuit is compact enough to include nearly 100 channels and the signal generator on 3 × 3mm die.

Neural recordings from high-density microelectrode arrays implanted in the cortex require time-frequency domain processing to alleviate the data telemetry bottlenecks of bandwidth and power. Our previous work has shown that the energy compaction capability of the discrete wavelet transform (DWT) offers a practical data compression solution that faithfully preserves the information in the neural signals. This paper presents a complete compression system including both lossy and lossless compression schemes, namely the DWT and run length encoding. Performance tradeoffs and key design decisions for implantable applications are analyzed. A 32-channel, 4-level version of the circuit is presented. Custom designed in 0.5 mum CMOS, occupying only 5.75 mm2and consuming 3mW of power (95 muW per channel at 25Ks/sec), the implantable compression circuit is well suited for intra-cortical neural interface applications.

Many emerging micro/nano sensor interfaces suitable for microsystem integration produce a change of impedance that must be monitored over a broad frequency range. This paper introduces a mixed-signal integrated circuit that can extract and digitize the real and imaginary components of a sensor's impedance response. The ultra compact size of this circuit enables each element in a multi-channel sensor array microsystem to have its own individual readout channel, permitting simultaneous readout and digitization of a high density sensor array. The circuit was fabricated in 0.5 mum CMOS and occupies only 0.045 mm2per cell. With a 3.3 V supply, each cell consumes only 5.2 muW at a typical 200 kHZ sampling frequency. For a 3 mm by 3 mm die, this circuit can be instantiated well over 100 times, which is sufficient for the needs of many anticipated sensor array microsystems.

Compact signal generators are a necessary component for many biomedical and chemical sensor microsystems. This paper presents a signal generator with precise digital frequency control that is 62% smaller the previous designs. The signal generator can produce analog sine waves and digital cosine waves from 4.8 Hz to 39 kHz with a SFDR greater than 99 dB. In a 0.5 µm CMOS process the total signal generator area is 361µm × 1048µm.

Chemiresistive (CR) sensors and sensor arrays coated with thiolate-monolayer-protected gold nanoparticle (MPN) interfaces show great promise for high-sensitivity multi-vapor analysis but suffer from process variation and drift in baseline values. This paper describes a new readout circuit that cancels baseline resistance and compensates for baseline drift to achieve ppm resolution. Requiring only 5100 mum in a 0.5 mum CMOS process, the circuit is well suited for high density on-chip CR sensor arrays. The resulting CR array microsystem introduces a valuable tool for monitoring environmental hazards including explosive compounds.

Discrete wavelet transform (DWT) has been shown to provide exceptionally efficient data compression for neural records. This paper describes an area-power minimized hardware implementation of the lifting scheme for multi-level, multi-channel DWT. Performance tradeoffs and key design decisions for implantable neuroprosthetics are analyzed. A 32-channel, 4-level version of the circuit has been custom designed in 0.18μm CMOS and occupies only 0.16mm2.

As nanostructured bioelectronic interfaces continue to evolve for sensor applications, new readout circuits are needed to harness their capabilities. This paper presents a single-chip amperometric readout circuit and electrode array system suitable for bioelectrochemical measurements. The chip features a CMOS potentiostat with high resolution, range-programmable current readout and electrochemical cell potential drive circuitry, which can perform on-chip chronoamperometry and cyclic voltammetry assays. Through post-CMOS fabrication, the surface of the chip is prepared with an array of electrodes suitable for formation of bioelectronic interfaces and on-chip bioelectrochemical measurements. The 3×3mm2chip nominally hosts a 4×4 working electrode array and supports amperometric outputs ranging from 10pA to 10μA with sub-pA resolution.

A receiving signal strength indicator (RSSI) built with transconductance amplifiers is presented. The RSSI achieves high tolerance to process variations by utilizing the unique nature of branch currents in a transconductance amplifier. These branch currents are used to implement a current-mode rectifier and amplitude clipping circuit that are tolerant of process variations. An on-chip offset control loop permits the entire RSSI to be realized with only one external component. In 0.18μm CMOS with a 1.8V supply, the RSSI draws 2.5mA and provides 80dB of offset suppression and more than 70dB of log-linear range with less than +/-2dB error due to process variation.

Reliable offset control is a vital feature of a zero-IF variable gain amplifier (VGA). Many traditional methods sacrifice speed (settling time) to achieve offset control by realizing a high pass feature in the signal path. This paper introduces a novel VGA architecture that is capable of controlling offset for all gain settings with a single offline calibration cycle. The VGA realizes very fast gain/startup settling time with only a slight increase in power and area requirements by eliminating the need for a high pass filter. The VGA circuit draws 3mA and has a gain range of 0/spl sim/42dB with 1dB steps. It has a 10MHz bandwidth at 42dB, has a gain settling time of 140nsec, and controls output offset within 2.2mV.

In a superscalar processor, select logic within the critical path of the instruction queue has become a performance bottleneck. This paper presents a high speed, two-level, hybrid select logic for wide-issue processors. The first level reduces delay by performing parallel age-based selection, and final arbitration is achieved in the second level with simple position-based select logic. The hybrid select logic circuits were implemented in dynamic logic on IBM 0.13/spl mu/m technology. Simulation shows 36% reduction in delay with less than 1% IPC degradation compared to the conventional design.

The instruction queue is a critical component and performance bottleneck in superscalar microprocessors. Conventional designs use physical register identifiers to wake up instructions. This paper proposes decoupling the tags for instruction wakeup from the tags for physical register access, thus increasing the design space of the instruction queue by encoding its operand tags. Two coding methods have been developed. One uses a linear code to increase the Hamming distance between tags, reducing the tag match delay by more than 50% and achieving 12% improvement in the total wakeup/select delay for TSMC 0.18mm technology at 1.8v. The second method uses one-hot code to encode the operand tag, removing the tag OR and tag read operations from the wakeup/select loop. For a 32-entry instruction queue, 15% reduction in the wakeup/select loop has been achieved. Furthermore, one-hot code also removes the dissipation-on-mismatch in the wakeup logic, significantly reducing the dynamic power consumption of the instruction queue.