Introduction

Lowering the costs of healthcare and increasing its accessibility is acritical need of today’s society. Miniaturized electronic sensors area possible way to both improve healthcare and lower the cost of medicaldiagnostics. Their small size and portability can lead to integration intopersonalized diagnostics tools and emergency care. In addition, faster,smaller and more efficient sensors can greatly impact chemical andbiological safety.

The rapid progress demonstrated in the computer industry and in genomics isstirring interest for growth and applications in healthcare and safety. Therelentless progress and developments in microfabrication predicted byMoore’s law, which forecast that the number of transistors on anintegrated circuit would double every two years, has until now beencontinually producing faster, cheaper, and smaller consumer electronics. Ananalogous exponential growth in DNA sequencing is even faster thanMoore’s law. Ten years ago, it would have taken many months tosequence a human genome. Today, the same task can be completed within oneday. This impressive progress is possible owing to innovative applicationsof microfabrication technologies. If this can be applied to healthcare, itcould stimulate a similar evolution, with applications such as early stagedetection of biological infection outbreaks and chemical hazards, whichcould mitigate epidemics tremendously.

The conventional brain–machine interface (BMI) concept is based on anelectrical circuit that includes electrodes for sensing and/orstimulation. In most cases, electrodes are in contact with biologicaltissue, while the electrical circuit can be positioned on the backside ofelectrodes or connected through a wireless link [1–3]. In both cases,a tremendous amount of work is done to design a high-sensitivity circuit todetect a few millivolts related to electrical signal propagation in thebrain [4, 5].

With recent advances in microfabrication and processing, more advancedsystems can be considered to build a BMI [6–8]. Following these majoradvances, the highly integrated lab-on-a-chip (LoC) emerged, which led tothe design of compact LoC in the size range of a few millimeters [9, 10].LoC has become attractive as an advanced brain–machine interface anda promising approach for future new brain discoveries.

Conventional brain–machine interfaces: deep brain electrodestimulation

There are two major operations that are handled by BMI, which are stimulationand sensing. Conventional BMIs are deep brain electrodes (DBE) or patchelectrodes [11, 12] used to detect the electrical activity of the brain. Inthe stimulation mode, DBE applies voltages or currents to tissue to generateaction potentials (AP). In the sensingmode electrodes are used to detect thepropagation of AP and brain electrical activity.

As stated in the previous chapter, Wolfgang Göpel wrote,“Bioelectronics is aimed at the direct coupling of biomolecularfunction units of high molecular weight and extremely complicated molecularstructure with electronic or optical transducer devices” [1]. Thestatement emphasizes two key points: the “direct coupling” andthe “high molecular weight”. This leads us now to developbioelectronic circuitries by moving from building blocks, or molecularcomponents, that involve high-molecular-weight biological molecules(typically proteins) by providing electrical contact at the nanoscale, inorder to address direct coupling. The first problem to solve is thereforethat of a nanoscale electrical contact.

During the last 15 years, several fabrication methods have been proposed toobtain a nanogap [2], based on the original work of Morpurgo etal. [3]. Among the possible solutions, a very effectivestrategy is to create an extremely tiny disconnection along a conductingwire in order to accommodate a biological molecule in between (Figure 2.1).The idea is to erode the electrical wire laterally in order to reduce itssize until one obtains a nanoscale conductor that becomes disconnected at acertain stage of the erosion. Morpurgo et al. [3]demonstrated that the wire’s conductivity becomes quantized in themoments immediately before the electrical connection breaks because thewavelength of the Schrödinger function associated with electronsbecomes comparable to the lateral size of the wire. We can therefore controlthe breaking of the connection by monitoring this quantum current. Now, wecan obtain a nanoscale interruption in the wire by stopping the erosionimmediately after the appearance of quantum states. If electrochemicaletching provides the erosion, then it is easy to control the process bymeans of a feedback system driven by the wire conductivity. Experimentalresults show that gaps can be obtained in the interrupted wire with sizesdown to 20 nm [3]. Considering now the typical size ofmetalloproteins, close to 5 nm [4], or of antibodies, up to15 nm [5], we can see that gaps of 20 nm are too large forquantum tunneling from the Fermi level in the metal to the LUMO (lowestunoccupied molecular orbital) in the protein [6].

Most probably, the term “lab-on-a-chip” came into the publicdomain with a nice publication that appeared in ScientificAmerican in 2007. That paper focused on recent developments intiny and portable chips for rapid testing in blood samples for pathogens orbiological weapons [1]. The article emphasized microfluidics as a keyfeature of the chip in order to make the lab-on-a-chip practical. The use ofair pressure or electricity is presented as the way to move and preciselymanipulate microscopic droplets through tiny chemical-reaction chambers [1].This sentence shows us the intimate connection between lab-on-a-chip andbioelectronics: we need electronics for precisely manipulating microscopicdroplets that contain pathogens or biological weapons. More than that, wealso need on board a quantitative technique for quantifying the pathogens orthe toxic agents. Therefore, biosensors are immediately involved. Figure35.1 schematically shows this deep interconnection addressed by modernsystems: a platform on a silicon chip that hosts a complex network ofmicrofluidic chambers and channels operating differently, and intimatelyintegrated with a complex microelectronic system. The whole system presentsloading ports for injection of samples and reagents. The loaded materialsare then mixed and pushed to the first reaction chamber by compressed air(as in point 1 of Figure 35.1) or by electrical or magnetic fields. Thereaction chamber performs thermal cycles to assure PCR (polymerase chainreaction) amplification (point 2 in Figure 35.1). In principle, a similarreaction chamber in the chip may perform other chemical reactions requiredby the assay (e.g. combination with another molecule to extract the targetmolecule from a complex). Then the reagent products are further moved intoanother reaction chamber to let a second reaction occur (point 3). Thesecond reaction may be required by the detection method. For example, areaction with a label (e.g. gold nanoparticles or fluorescent dyes) may beprovided in the case of labelled detection. Finally, the marked productsarrive in the analysis chamber, and the products are detected (point 4). Inthe example of Figure 35.1, the detection is provided by a series of diodesthat detect the pathogens as well as by electrophoretic tests.

This chapter explores electrical stimulation of excitable biologicalentities. It provides a brief history of electrical stimulation, stimulationparameters, theory of electrode interface impedance, and types ofstimulation. Focus is on the technology of charge balancing, power losses,voltage compliance, and associated hardware complexities of electricalstimulators. Stimulus artifacts and inductive power delivery are alsodiscussed, and are often used with stimulators for applications in neuralprostheses and therapeutics [1, 2].

Introduction

Electrical stimulation is an important and popular method in which injectedcharges are mainly responsible for excitation or inhibition of neural ormotor structures [3–5]. It can be utilized for treatment of diseasesand restoration of dysfunctional organs, such as for the brain [6, 7],retina [8, 9], cochlea [10], or peripheral nerve [11]. The electricalstimulation process is accomplished by using conductive electrodes ofspecific size and shape, which are placed at desired biological sites ofinterest. Coulombic charge is injected through such electrodes by means ofconstant current or constant voltage pulses. Charge injection accuracy,power loss minimization, output voltage compliance, and miniaturization areimportant factors that can enhance practical performance in electronicstimulators [12]. In the following subsections we provide a brief historyand introduce technical parameters associated with electricalstimulation.

Introduction

The notion of creating artificial vision using visual prostheses has beenwell represented though science fiction literature and films. When we thinkof retinal prostheses, we immediately think of fictional characters like TheTerminator scanning across a bar to assess patrons for appropriately fittingclothing, or Star Trek’s Geordi La Forge with his VISOR, a visualinstrument and sensory organ replacement placed across his eyes and attachedinto his temples to provide him with vision. Such devices are no longerfarfetched. In the past 20 years, significant research has been undertakenacross the globe in the race for a “Bionic Eye”. Advances inBionic Eye research have come from improvements in the design andfabrication of multielectrode arrays (MEAs) for medical applications. MEAsare already commonplace in medicine with use in applications such as thecochlear device, cardiac pacemakers, and deep brain stimulators whereinterfacing with neuronal cell populations is required.

The use of MEAs for vision prostheses is currently of significant interest.For the most part, retinal prostheses have dominated the research landscapeowing to the ease of access and direct contact to the retinal ganglion nervecells. However, MEAs are also in use for direct stimulation into the opticnerve [1]. Retinal prostheses bypass the damaged photoreceptor cells withinthe retina and instead replace the degenerate retina with electricalstimulation to the nerve cells. Using electrical stimulation, stimulatedretinal ganglion cells have been shown to elicit a percept in the form of aphosphene in blind patients [2–6]. Accordingly, the two diseasescommonly linked to the justification for Bionic Eye research are age-relatedmacular degeneration (AMD) and retinitis pigmentosa (RP), diseases whichlead to progressive loss of photoreceptor cells and diseases where thepatient has had previous vision and thus exhibits prior visual-brainpathways. At present, there has been no reliable cure for any of the retinaldiseases that target the photoreceptor cells, and thus the development ofprosthetic devices is a viable clinical treatment option [7–9].

Introduction to motor prosthetics

Fritsch and Hitzig first discovered the motor cortex in 1870 [1], althoughthe best-known experimental mapping of the motor cortex dates back toPenfield’s experiments in 1937 [2] using electrical stimulation toactivate muscle groups in patients undergoing surgery for epilepsy. It wasnot until the 1980s, over 100 years since the discovery of the motor cortex,that population coding [3] was proposed and thus the beginnings of decodingneural signals in the motor cortex into their corresponding motorfunction.

In 1998 the first human was implanted with a brain–machine interface(BMI) of high enough quality to simulate movement and demonstratedtwo-dimensional control of a mouse cursor [4],[5]. Since then, there hasbeen an explosion of demonstrations of motor prosthetic control of computercursors and robotic arms by both primates and humans. In the last year, thesame group demonstrated robotic arm control with four degrees of freedom ina tetraplegic patient [6]. These demonstrations mark a significant step inbringing BMIs from the research arena to viable medical devices, but anumber of technological hurdles must still be overcome to make this areality.

A simplified diagram of a BMI system is exemplified in Figure 29.1. A fullBMI system involves a recording device to take signals directly from themotor cortex.

Abstract

In the past few years biosensing concepts based on organic field-effecttransistors (OFETs) have attracted more and more attention. Here organicelectronics benefit especially from the fact that solution-processableorganic thin films can be used in flexible and disposable sensors.Additionally, the outstanding biocompatibility of many organic materialsallows the use of organic sensing devices for in-vivoapplications and permits the design of biodegradable sensors.

Starting from the basic principles of organic thin-film transistors, thischapter will present the state of the art of biosensing approaches based onOFETs, either back-gated or electrolyte-gated, focusing in particular ondifferent functionalization methods to achieve a selective response of theOFET towards biologically relevant molecules. We present recently publishedapplications of organic thin-film transistors ranging from the detection ofbiomolecules such as DNA, proteins, and enzymes to the sensing andstimulation of electrical activity potentials of neurons. The sensingmechanism and the influence of the Debye screening length on the detectionof biomolecules will be discussed.

Background and introduction

Early and correct diagnosis of diseases plays a crucial role in modernmedicine. For several decades researchers have been working on the discoveryof biomarkers, molecular indicators giving early information about variousdiseases. For detection of the increasing number of biomarkers, low-cost,fast, and reliable methods are required.

Recent advances in digital microfluidics have led to tremendous interest inminiaturized lab-on-a-chip devices for biochemical analysis. Synthesis toolshave also emerged for the automated design of lab-on-a-chip from thespecifications of laboratory protocols. However, none of these toolsconsiders control flow or addresses the problem of recovering from fluidicerrors that can occur during on-chip bioassay execution. We present asynthesis method that incorporates control paths and an error-recoverymechanism in the design of a cyberphysical digital microfluidiclab-on-a-chip. A microcontroller coordinates the implementation of thecontrol-flow-based bioassay by intercepting the synthesis results that aremapped to the software programs. Real-life bioassay applications are used ascase studies to evaluate the proposed design method. For a representativeprotein assay, compared with a baseline chip design, the biochip with acontrol path can reduce the completion time by 30% when errors occur duringthe implementation of the bioassay. A fabricated biochip is also used todemonstrate cyberphysical error recovery in a laboratory setting.

Introduction

Microfluidic biochips have now come of age, with applications to biomolecularrecognition for high-throughput DNA sequencing, immunoassays, andpoint-of-care clinical diagnostics [1]. In particular, digital microfluidicbiochips, which use electrowetting-on-dielectric to manipulate discretedroplets (or “packets of biochemical payload”) of picolitervolumes under clock control, are especially promising [2].

Characterization of molecular electronic transport is an active part of theresearch field in nanotechnology. The main underlying idea is to use singlemolecules as active elements in nanodevices [1]. As a consequence, theproper fabrication of a molecule–electrode contact is a crucial issue[2],[3] and several applications can be envisioned. For example, thevariation of the electrical conduction of metal–molecule–metaljunctions can be used in biosensing for electrochemical detection ofdifferent crucial biomarkers. Possible applications include biomedicaldiagnostics and the monitoring of biological systems. In particular, thedetection of single proteins might become the starting point for monitoringdrugs, developing clean energy systems, fabricating bio-optoelectronictransistors, and developing other innovative devices and systems.

The fabrication of nanogaps

Nanogap electrodes (NGEs, defined as a pair of electrodes separated by ananometer-sized gap) are fundamental tools for characterizing the electricproperties of material at the nanometer scale, or even at the molecularscale. They are also important building blocks for the fabrication ofnanometer-sized devices and circuits.

Molecular-based devices possess unique advantages for electronic applicationswith respect to conventional components [4], such as lower cost, lower powerdissipation and higher efficiency. Specific molecules can be not onlyrecognized, but also self-assembled on such NGEs, thus leading to elaborategeometries for the study of distinct optical and electronic properties.

Realization of bio-inspired computational systems demands elements that willbe used for both memorizing and processing of the information, as occurs inthe brain. In this chapter we will consider organic memristive devices– elements that were designed and constructed for mimicking the mostimportant property of synapses, responsible for so-called Hebbian learning.The organization of the device and its important properties are considered.Its utilization in generators of auto-oscillation generators and in logicelements with memory is also considered. Finally, stochastic networks oforganic memristive devices have demonstrated several similarities withlearning of brains of animals and humans.

Introduction

A significant difference in the architecture of computers and the brain isthat in the computer the memory and the processor are different devices withno influence on each other. The information in this case plays a passiverole – it can be recorded, accessed, canceled, but it does not varyproperties of the system. In the brain, in contrast, the same elements areused for both memorizing and processing of the information. Thisarchitecture allows learning of the system at a hardware level. Informationbegins to play an active role: it is not only recorded, but it variesconnections within the processor, preparing it for more effective resolvingof similar tasks in the future. The other essential difference between thebrain and computer is the fact that in the brain we have highly parallelinformation processing. This is why it is much more effective for sometasks, such as, for example, the recognition and classification ofobjects.

Introduction

Biomolecular detection is crucial from various perspectives, such as qualitycontrol of our food and water, identification of biological terroristagents, and diagnosis of diseases. Early detection of disease is importantfor effective treatment and for prognostic assessment of diseaseprogression; in addition, the trend of ageing societies leads to anincreasing requirement for biomarker diagnoses for personalized healthcaremonitoring. This results in more stress on the social healthcare system [1,2]. As a consequence, researchers have focused on developing biomoleculardetection devices and systems. Over the past decade, emerging methods toaddress the above needs have bloomed because of developments inmicro/nanotechnologies. To enhance throughputs and reduce costs,moreover, these detection devices and systems are evolving from label-basedto label-free technologies.

Traditionally, label-based molecular diagnosis techniques have been used as auseful fundamental concept for the detection of potential disease biomarkersor pathogen nucleic acids. In general, the detection signal comes from theusage of a specific tag for a target molecule. The tags can be conventionalfluorescent dyes or radioisotopes. To fulfill the requirements of differentapplications, a number of conventional label-based techniques, such aspolymerase chain reaction (PCR), DNA or protein microarrays, andenzyme-linked immunosorbent assay (ELISA), have been developed andimplemented. Some of them have been used to form a versatile platform formany diverse applications with promising results and represent the goldstandards of biomedical diagnosis [3–5]. However, these techniquesrequire trained staff and expensive equipment, and are time-consuming.Moreover, the detection of such low-abundance biomarkers in biologicalfluids (e.g. blood, urine, saliva) requires large quantities of the sampleand complicated sample preparation. Consequently, these label-basedtechniques encounter problems of cost-effectiveness and throughput undermodern circumstances.

From their discovery, CNTs have increasingly attracted interest because oftheir peculiar electrical, mechanical, and chemical properties. In 1991,Sumio Iijima first observed and described in detail the atomic arrangementof this new type of carbon structure [1]. By a technique used for fullerenesynthesis, he produced needle-like tubes at the cathode of an arc-dischargeevaporator. From that time, carbon nanotubes have been used for manyapplications and represent one of the most typical building blocks used innanotechnology. Their peculiarities include unique properties of fieldemission and electronic transport, higher mechanical strength with respectto other materials, and interesting chemical features.

The use of CNTs has recently gained momentum in the development ofelectrochemical biosensors, since their utilization can create devices withenhanced sensitivity and detection limit capable of detecting compounds inconcentrations comparable to those present in the human body.

This chapter will review the most important features of carbon nanotubes, andpresent an example in which their application can enhance the detection ofdrugs and metabolites relevant in personalized medicine: P450 biosensors fortherapeutic drug monitoring.

Overview

Carbon is a very interesting element, since it can assume several stablemolecular structures. Any molecule entirely composed of carbon is called afullerene.

Knee: functional anatomy I

The knee is the largest and most complex joint in the body. It is an articulation among the distal femur, proximal tibia and patella, surrounded by a soft tissue envelope of capsule, ligaments, tendons and muscles. The specific anatomical features and interactions of these structures determine the stability and movements of the knee joint.

Tibiofemoral joint

The medial tibial plateau is larger than the lateral tibial plateau. It also has a concave depression whereas the lateral tibial plateau has a convex surface. Both plateaux have a posterior tilt of about 7°, but the joint surface tilt is minimised to 3° by the menisci. The plateaux also have a 3° varus tilt, i.e. the joint surface is slopped medially.

On the femoral side, the medial condyle is larger, more curved and extends more distally than the lateral condyle.

This asymmetry between the medial and lateral compartments of the knee determines the relative mobility of each compartment. The concave medial tibial plateau offers more constraint and the convex lateral tibial plateau offers less constraint to the movements of their respective femoral condyles. The posterior and varus tilts of the tibial plateaux have implication for the alignment of bone cuts for total knee replacement.