WO2025212114A1 - Systems and methods for adaptive feedback control for drug concentrations management - Google Patents

Abstract

Methods and systems that use adaptive feedback control to control the in vivo concentrations of target molecules or ions accurately and rapidly by estimating the pharmacokinetics of the individual subject at site of control are described. The systems can estimate the pharmacokinetics of the molecules or ions under control without prior pharmacokinetic knowledge of the subject's specific physiology or the pharmacokinetics of the molecules or ions under control, ensuring that the resulting control is rapid and accurate. The adaptive feedback control systems can rapidly achieve and accurately maintain a predefined concentration in the tissue that is the site of action of the molecules or ions.

Description

SYSTEMS AND METHODS FOR ADAPTIVE FEEDBACK CONTROL FOR DRUG CONCENTRATIONS MANAGEMENT

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] The current application claims the priority to U.S. Provisional Patent Application No. 63/575,313 entitled “Adaptive Feedback Control for Drug Concentrations Management” filed April 5, 2024. The disclosure of U.S. Provisional Patent Application No. 63/575,313 is hereby incorporated by reference in its entirety for all purposes.

GOVERNMENT SPONSORED RESEARCH

[0002] This invention was made with government support under AI145206 awarded by the National Institutes of Health. The government has certain rights in the invention.

FIELD OF THE INVENTION

[0003] The current disclosure is directed to systems and methods for measuring, estimating, and controlling the concentrations of drugs, metabolites, hormones, and other biologically and medically important molecules and ions; and more particularly to adaptive feedback control systems for controlling the concentrations of such species in a subject’s body.

BACKGROUND

[0004] Pharmacokinetic models are models describing distribution and elimination kinetics of drugs and other molecules and ions within the body. The models describe the relationship between the dose administered to an individual and the resulting concentrations inside various body compartments. Such models can be an important part of the proper dosing of drugs, the treatment of diseases, and of the drug discovery and development processes. BRIEF SUMMARY

[0005] Systems and methods for estimating and controlling the concentrations of various molecules and ions in the body using adaptive feedback controls are described.

[0006] Some embodiments include an adaptive feedback control system, comprising:

• an electrochemical aptamer based (EAB) sensor, comprising: a working electrode, wherein the working electrode comprises a recognition element modified with a redox molecule; wherein the recognition element specifically binds to a substance and is configured to undergo a conformational change upon binding with the molecule or ion; wherein a current readout from the EAB sensor at a site of control reflects a concentration of the substance;

• a delivery means; wherein the delivery means is configured to deliver the substance to a site of delivery; and

• a feedback controller; wherein the feedback controller is configured to modulate a delivery rate of the delivery means; wherein the feedback controller predicts pharmacokinetics of the substance at the site of control in response to the delivery rate and the concentration of the substance such that the feedback controller adjusts the delivery rate in real time to achieve a predefined concentration of the substance at the site of control.

[0007] In some embodiments, the site of control and the site of delivery are different.

[0008] In some embodiments, the site of control is a body part or a region of a subject.

[0009] In some embodiments, the site of control is a site of drug action.

[0010] In some embodiments, the site of control is the brain of a subject, and the site of delivery is selected from the group consisting of: a vein, an artery, a muscle, an intraperitoneal space, and a subcutaneous space.

[0011] In some embodiments, the substance comprises a plurality of molecules, or a plurality of ions.

[0012] In some embodiments, the feedback controller implements an initial delivery profile by injecting the substance at a feasible injection rate until the delivery means achieves a predefined intermediate concentration of the substance; after the initial injection, the feedback controller reduces the injection rate by a pre-defined amount until the delivery achieves a second predefined, intermediate concentration; wherein the injection rate is clinically feasible, or technically feasible, or physiologically feasible, or animal-welfare appropriate.

[0013] In some embodiments, the pre-defined amount is by half.

[0014] In some embodiments, the predefined intermediate concentration is between 30% and 40% of a predefined desired concentration of the substance at the site of control; wherein the second predefined intermediate concentration is between 60% and 70% of the predefined desired concentration of the substance at the site of control.

[0015] In some embodiments, after the feedback controller reaches a threshold value, the feedback controller switches from the initial injection profile to an optimal injection profile.

[0016] In some embodiments, the optimal injection profile uses an optimal control input calculated by solving a mathematical equation at a first time step; at a time step after the first time step, the feedback controller uses a measured concentration and a corresponding injection rate to replace the optimal control input such that the feedback controller updates and adjusts the delivery rate to an optimal delivery rate at each time step to achieve the predefined concentration at the site of control.

[0017] In some embodiments, the feedback controller does not use population data- based pharmacokinetics of the substance.

[0018] In some embodiments, the feedback controller does not use prior measurements of pharmacokinetics of the substance.

[0019] In some embodiments, the feedback controller is not initialized before a start of control.

[0020] In some embodiments, the site of control is a body part of a subject and the feedback controller is initialized before a start of control using population pharmacokinetic data to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics. [0021] In some embodiments, the site of control is a body part of a subject and the feedback controller is initialized before a start of control using foreknowledge or an estimate of a subject’s specific pharmacokinetics to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

[0022] In some embodiments, the delivery means is a catheter or a syringe.

[0023] In some embodiments, the recognition element is a nucleic acid.

[0024] In some embodiments, the recognition element is a DNA aptamer.

[0025] In some embodiments, the working electrode is a wire, a rod, a planar electrode, a needle, or a microneedle.

[0026] In some embodiments, the conformational change alters an electron transfer rate between the redox molecule and a surface of the working electrode.

[0027] In some embodiments, the substance is a drug molecule selected from the group consisting of: a small molecule drug molecule, an amino acid-based drug molecule, a protein-based drug molecule, a lipid-based drug, a polysaccharide-based drug and a nuclear acid-based drug molecule.

[0028] In some embodiments, the substance is a non-drug molecule selected from the group consisting of: a hormone, an enzyme, an antibody, a cytokine, and a metabolite.

[0029] In some embodiments, the substance is an ion selected from the group consisting of sodium, potassium, lithium, calcium, magnesium, and chloride.

[0030] In some embodiments, the substance is procaine, the recognition element is a procaine-binding DNA aptamer, the redox molecule is methylene blue, the site of delivery is a peripheral tissue, and the site of control is the brain.

[0031] In some embodiments, the EAB sensor further comprises a counter and reference electrode pair, or a counter electrode and a separate reference electrode.

[0032] Some embodiments include a method for adaptive feedback control, comprising:

• delivering a substance to a cite of delivery;

• measuring a concentration of the substance using at least one EAB sensor; wherein the at least one EAB sensor comprises a working electrode comprising a recognition element modified with a redox molecule; wherein the recognition element specifically binds to the substance and is configured to undergo a conformational change upon binding; wherein a current readout from the EAB sensor at a site of control reflects a concentration of the substance; and

• controlling a delivery rate via a feedback controller; wherein the feedback controller predicts pharmacokinetics of the substance at the site of control in response to the delivery rate and the concentration of the substance such that the feedback controller adjusts the delivery rate in real time to achieve a predefined concentration of the substance at the site of control.

[0033] In some embodiments, the site of control and the site of delivery are different.

[0034] In some embodiments, the site of control is a body part of a subject.

[0035] In some embodiments, the site of control is a site of drug action.

[0036] In some embodiments, the site of control is the brain of a subject, and the site of delivery is selected from the group consisting of: a vein, an artery, a muscle, an intraperitoneal space, a subcutaneous space, and a peripheral tissue of the subject.

[0037] In some embodiments, the feedback controller implements an initial injection profile by delivering the substance at a feasible injection rate until the delivery achieves a predefined intermediate concentration; after the initial injection, the feedback controller reduces the injection rate until the delivery achieves a second predefined, intermediate concentration; wherein the injection rate is clinically feasible, or technically feasible, or animal-welfare appropriate.

[0038] In some embodiments, the predefined intermediate concentration is between 30% and 40% of a predefined desired concentration at the site of control; wherein the second predefined intermediate concentration is between 60% and 70% of the predefined desired concentration at the site of control.

[0039] In some embodiments, after the feedback controller reaches a threshold value, the feedback controller switches from the initial injection profile to an optimal injection profile. [0040] In some embodiments, the optimal injection profile uses an optimal control input calculated by solving a mathematical equation at a first time step; at a time step after the first time step, the feedback controller uses a measured concentration and a corresponding injection rate to replace the optimal control input such that the feedback controller updates and adjusts the delivery rate to an optimal delivery rate at each time step to achieve predefined concentration of the substance at the site of control.

[0041] In some embodiments, the feedback controller does not use population data- based pharmacokinetics of the substance.

[0042] In some embodiments, the feedback controller is not initialized before a start of control.

[0043] In some embodiments, the site of control is a body part of a subject and the feedback controller is initialized before a start of control using population pharmacokinetic data to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

[0044] In some embodiments, the site of control is a body part of a subject and the feedback controller is initialized before a start of control using foreknowledge or estimates of a subject’s specific pharmacokinetics to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

[0045] In some embodiments, the feedback controller does not use prior measurements of pharmacokinetics of the substance.

[0046] In some embodiments, the delivery is via a catheter or a syringe.

[0047] In some embodiments, the recognition element is a nucleic acid.

[0048] In some embodiments, the recognition element is a DNA aptamer.

[0049] In some embodiments, the working electrode is a wire, a rod, a planar electrode, a needle, or a microneedle.

[0050] In some embodiments, the conformational change alters an electron transfer rate between the redox molecule and a surface of the working electrode.

[0051] In some embodiments, the at least one EAB sensor further comprises a counter and reference electrode pair, or a counter electrode and a separate reference electrode. [0052] In some embodiments, the substance comprises a plurality of molecules, or a plurality of ions.

[0053] In some embodiments, the substance is a drug molecule selected from the group consisting of: a small molecule drug molecule, an amino acid-based drug molecule, a protein-based drug molecule, a lipid-based drug molecule, a polysaccharide-based drug molecule and a nuclear acid-based drug molecule.

[0054] In some embodiments, the substance is a non-drug molecule selected from the group consisting of: a hormone, an enzyme, an antibody, a cytokine, and a metabolite.

[0055] In some embodiments, the substance is an ion selected from the group consisting of sodium, potassium, lithium, calcium, magnesium, and chloride.

[0056] In some embodiments, the substance is procaine, the recognition element is a procaine-binding DNA aptamer, the redox molecule is methylene blue, the site of delivery is a peripheral tissue, and the site of control is the brain.

[0057] Additional embodiments and features are set forth in part in the description that follows, and in part will become apparent to those skilled in the art upon examination of the specification or may be learned by the practice of the disclosed subject matter. A further understanding of the nature and advantages of the present disclosure may be realized by reference to the remaining portions of the specification and the drawings, which form part of this disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The description will be more fully understood with reference to the following figures, which are presented as example embodiments of the invention and should not be construed as a complete recitation of the scope of the invention, wherein:

[0059] Figure 1A illustrates electrochemical aptamer-based (EAB) sensors comprising a gold working electrode and a stainless-steel counter/pseudoreference electrode in accordance with an embodiment. The working electrode is functionalized by chemically attaching a redox-reporter-modified aptamer (an oligonucleotide sequence selected in vitro to bind to a specific target) that undergoes a conformational change upon target binding.

[0060] Figure 1 B illustrates a change in peak current when the sensor is interrogated using square wave voltammetry in accordance with an embodiment.

[0061] Figure 1 C illustrates EAB sensors’ support of real-time concentration measurements, thus enabling closed-loop feedback-controlled delivery of specific molecules to the body in accordance with an embodiment. Controlling the concentration of a drug in the site of action, such as in-brain for anesthetic procaine, can be complicated by slow transport dynamics into the tissue and significant subject-to-subject variability in the transport dynamics from the site of drug delivery (in-vein) and action (in-brain). The use of adaptive feedback control in this disclosure can rapidly and accurately achieve the desired concentration of a specific molecule in the plasma or any other bodily compartment.

[0062] Figures 2A through 2C show a PID controller that was designed using information derived from population pharmacokinetics is used to perform feedback control over in-brain drug concentrations in accordance with an embodiment. Figure 2A uses uncontrolled, 20-min drug infusions to experimentally determine the pharmacokinetics of 7 animals. Using the pharmacokinetics determined for each of the 7 individual animals, a PID controller is designed to achieve the targeted concentration in all 7 without significant overshooting in any of them.

[0063] Figure 2B shows employing the PID controller in an 8^th animal that is not part of the training set, the controller reaches and accurately maintains (to a root-mean-squared (RMSD) error of 18 |iM) the desired (i.e., set point) 100 |iM drug concentration (the dashed line). However, because this PID controller does not accurately reflect the pharmacokinetics of any individual animal, it needs to be tuned rather conservatively to avoid overshooting. Because of this, it takes nearly 1 h to reach its desired set point.

[0064] Figure 2C shows subject-to-subject pharmacokinetic variability can also lead to unexpected behavior when using a population-based controller. Figure 2C shows the simulated responses of rats from previous experiments during the rising phase of the actual experiment, (highlighted in 2B) when the difference is clearest. The trend line in the feedback control experiment (blue) at first reacts more slowly and then ramps up more quickly than all the previously observed behavior. The datasets are plotted with the same colors as the simulated responses in 2A.

[0065] Figure 3A shows that an autoregressive moving average (ARMA) model of order (2, 1 ) (as opposed to a higher order or more complex model) is sufficient to model the in-brain pharmacokinetics of procaine by showing the compatibility of such a model with the previous (PID controller) in-brain feedback control experiment in accordance with an embodiment. Specifically, at about 60 min (vertical dashed line), the ARMA (2, 1 ) model (light-blue line) is fitted to the prior 100 measurements, and this fitted model is used to predict the concentrations expected over the next 50 measurements (magenta line). Notably, the large majority of the measured concentrations (dots) fell within the estimated 95% confidence intervals (defined by the magenta dotted lines) of the ARMA (2,1 ) model prediction.

[0066] Figure 3B shows that the model continues to accurately predict the concentrations observed 50 measurements ahead throughout the entire PID control experiment in accordance with an embodiment. Specifically, 78% of the observations fall within the estimated 95% confidence intervals around ARMA (2,1 ) model prediction (dotted lines), with the discrepancy between about 78% and 95% presumably arising due to fluctuations in the animal’s physiology and the noise levels of the sensor.

[0067] Figure 3C shows that the precision of the ARMA model’s predictions is greater for predictions that do not extend as far in the future in accordance with an embodiment. In Figure 3C, shown are the estimated 95% confidence limits for the model’s predictions forecast 1 (blue dotted line), 25 (green dotted line), or 50 (magenta dotted line) measurements ahead of the most recent measurement. The precision of these concentration predictions falls as the model predicts further into the future: the estimated 95% confidence intervals around the 1 measurement-ahead prediction capture 89% of the actual measurements, the 25 measurements ahead prediction captures 84%, and the 50 measurements ahead prediction captures 78%. [0068] Figure 4 shows a snapshot of the adaptive control algorithm at work in accordance with an embodiment. As its first step, the system identification algorithm identifies the pharmacokinetic model that best describes the recently observed data (the light blue line in the top panel). The drug delivery started by using a pre-determined “initialization injection” (the green line in the lower panel) which employs a pulse-like shape to provide maximum information regarding the individual animal’s pharmacokinetics while simultaneously approaching the desired concentration rapidly. After this initial injection profile is complete, the adaptive feedback controller takes control over the infusion rate (the purple line in the lower panel) and uses the optimal injection profile (the magenta lines in the lower panel) that, given the pharmacokinetic model, is predicted to drive the concentration to its targeted value as rapidly as possible (magenta line, upper panel). At each new measurement, the pharmacokinetic model is refined and/or updated, and the cycle is repeated.

[0069] Figure 5A shows that adaptive control rapidly achieves and precisely maintains its targeted concentration profile in accordance with an embodiment. The algorithm can achieve and remain at an in-brain procaine level of about 100 pM for about 1 h with an RMSD of 10.2 pM.

[0070] Figure 5B repeated the prior experiment in a second animal with a targeted concentration of about 200 pM in accordance with an embodiment. This concentration (with an RMSD value of 11 .3 pM) is held for about 1 h.

[0071] Figures 5C and 5D show that after establishing that the algorithm can maintain constant concentrations, the algorithm is used to achieve pre-defined, time varying concentration profiles in which the targeted concentration is stepped from high to low (Figure 5C) and from low to high (Figure 5D) in accordance with an embodiment. For each data set the point (green lines) is highlighted at which the initialization injection ends, and the adaptive controller takes over drug delivery. The trend seen in data is highlighted by plotting a smoothed curve (blue lines) obtained by passing the raw data (black dots) through first a Hampel filter over 10-minute interval and second moving average smoothing filter over a 5-minute interval. DETAILED DESCRIPTION

[0072] After considering this description it will be apparent to one skilled in the art how the invention is implemented in various alternative embodiments and alternative applications. However, although various embodiments of the present invention will be described herein, it is understood that these embodiments are presented by way of example only, and not limitation. As such, this description of various alternative embodiments should not be construed to limit the scope or breadth of the present invention. Furthermore, statements of advantages or other aspects apply to specific exemplary embodiments, and not necessarily to all embodiments, or indeed any embodiment covered by the claims.

[0073] Throughout the description and the claims of this specification the word “comprise” and variations of the word, such as “comprising” and “comprises” are not intended to exclude other additives, components, integers, or steps.

[0074] Reference throughout this specification to “one embodiment” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, appearances of the phrases “in one embodiment” or “in an embodiment” in various places throughout this specification are not necessarily all referring to the same embodiment, but may.

[0075] The term “subject” is used to refer to an animal (including a human and a nonhuman animal) to which the present invention may be applied.

[0076] The ability to perform feedback control over drug, hormone, metabolite, or protein concentrations in the tissues that are the site of action of these molecules could significantly impact biomedical research and clinical practice. Control over tissues other than the blood, however, is rendered difficult by the time-lags associated with intercompartmental drug transport. A population-based approach to performing such control was previously used, in which the control algorithm was optimized using the known population pharmacokinetics of the drug under control in the tissue of relevance. This approach, however, may fail to work properly in individuals whose pharmacokinetic response differs significantly from that of the original test population. To overcome the shortcomings of the previous approaches, many embodiments implement an adaptive feedback-control mechanism that can rapidly reach and accurately maintain the desired set point concentration without relying on any prior pharmacokinetic knowledge. In several embodiments, the adaptive feedback-control systems can achieve rapid attainment and accurate (root-mean-squared deviation 5 to 10%) maintenance of the concentration of the anesthetic procaine in the brains of live rats.

[0077] Electrochemical aptamer based (EAB) sensors can be used for monitoring drugs, metabolites, hormones, and proteins and other biologically important molecules and ions in the body continuously (i.e., at frequencies matched to physiological timescales) and/or in real-time (i.e., with lag times far shorter than physiological timescales). (See, e.g., Xiao, Y., et al., 2005, Angewandte Chemie, 117(34), 5592-5595; Arroyo-Curras, N., et al., 2018, ACS Sensors, 3(2), 360-366; the disclosures of which are incorporated by reference.) EAB sensors comprise a redox-reporter-modified aptamer (nucleic acidbased receptors selected in vitro to bind specific molecular targets) covalently attached to an electrode (Figure 1A). Target binding causes an alteration in the rate of electron transfer from the attached redox reporter (Figure 1 B). This signal transduction mechanism renders EAB sensors (i) reagentless, (ii) single-step and rapidly reversible, (iii) highly selective, and (iv) independent of the chemical or enzymatic reactivity of their target and thus generalizable. Building on this, EAB sensors have been reported against dozens of molecular targets, more than a half dozen of which have been measured in situ in live rats. These include second- or sub-second resolved, real-time measurements performed intravenously (i.e., in blood plasma), in the subcutaneous space (in interstitial fluid, tumors), in the brain (the cerebrospinal fluid of the lateral ventricles and hippocampus), and in solid tumors. (See, e.g., Arroyo-Curras, N., et al., 2018, ACS Pharmacology & Translational Science, 1(2), 110-118; Chamorro-Garcia, A., et al., 2023, ACS Sensors, 8(1 ), 150-157; Idili, A., et al., 2021 , Analytical Chemistry, 93(8), 4023-4032; Lin, S., et al., 2022, Science Advances, 8(38), eabq4539; Wu, Y., et al., 2022, Analytical Chemistry, 94(23), 8335-8345; Gerson, J., et al. , 2023, Science Advances, 9, 20; Wu, B., et al., 2023, Micromachines, 14, 323; J.W., Fu, K., et al., 2022, Sci Adv., 8, eabk2901 ; the disclosures of which are incorporated by reference.)

[0078] The real-time, in vivo concentration information provided by EAB sensors provides unique opportunities to perform closed-loop feedback control over the concentrations of drug and other biologically important molecules and ions in specific bodily compartments (Figure 1 C). As proof of concept, EAB sensors and a proportionalintegral-derivative (PID) feedback controller were used to achieve the high-precision control of plasma levels of the antibiotics tobramycin and vancomycin. (See, e.g., Dauphin-Ducharme, P., et al., 2019, ACS Sensors, 4(10), 2832-2837; the disclosure of which is incorporated by reference.) Specifically, this approach can be used to control intravenous infusions and can maintain constant plasma drug concentrations (to within better than ±20%) as well as plasma drug concentrations that accurately follow predefined, time-varying profiles. Such profiles mimic human pharmacokinetics in a rat. Another PI D-based feedback control can be used for intracranial concentrations of the anesthetic procaine using a sensor placed in the lateral ventricle of a rat and intravenous drug delivery. This ability to accurately control the molecular concentrations at specific sites in the body has the potential to impact the development of new drugs and the delivery of treatments to patients. For example, the ability to precisely control in vivo drug concentrations in animal models could improve the accuracy and efficiency of the preclinical stages of drug development by, for example, eliminating pharmacokinetic differences between animals as a confounding experimental variable. Such a technology could also improve the efficiency and the safety of pharmaceutical treatments by ensuring that drug concentrations remain within the therapeutic window, thus avoiding undertreatment and toxicity.

[0079] A benefit of the EAB sensor-enabled ability to perform feedback control in the brain and peripheral tissues is that it opens the door to controlling a drug’s concentrations in the tissues that are its site of action. Examples of this include (but are not limited to) the control over an anesthetic in the brain, a chemotherapeutic in a tumor, or an antibiotic at a site of infection. The slow intercompartmental transport of drugs, however, complicates feedback control in compartments other than the compartment to which the drug is being delivered to. That is, when the site of delivery of the molecule or ion, such as into the vein (plasma), and the site of measurement, such as in the brain, are different, the finite rate of transport between the two compartments renders feedback control more difficult. The magnitude of this obstacle can vary from depending on the physiological status of the subject, on the precise tissue site under control, and on the physicochemical and pharmacological properties of the substance (i.e., the molecule and/or ion) being delivered, renders it challenging to design simple and effective PID controllers. In response to this challenge, many embodiments implement an adaptive feedback control algorithm that models the site-specific pharmacokinetics of the target molecule during the initial phases of control autonomously and in real time, and then continues to update its model for the duration of the control, thus rapidly achieving and accurately maintaining a predefined concentration in the tissue that is the site of drug action.

[0080] Many embodiments can provide feedback control over various types of drugs including (but not limited to) small molecule drugs, large molecule drugs, amino acidbased drugs, protein-based drugs, nuclei acid-based drugs. The drugs can have a variety of sizes and/or molecular weights.

[0081] Many embodiments can provide feedback control over various types of biomolecules, such as (but not limited to) hormones, metabolites, proteins, enzymes, antibodies, and/or oligonucleotides.

[0082] Many embodiments can provide feedback control over various types of ions, such as (but not limited to) sodium, potassium, magnesium, and/or calcium.

[0083] In many embodiments, the delivery site of the substance and the site of control can be different. In certain embodiments, the site of control can be the brain of a subject. In certain embodiments, the delivery site can be any of a desired location for a subject such as (but not limited to) veins, arteries, muscles, intraperitoneal space, subcutaneous space, and peripheral tissues. [0084] In many embodiments, the EAB sensors can be used for measuring molecules and/or ions concentrations. EAB sensors can perform seconds and/or sub-seconds resolved measurements of multiple molecules and/or ions in situ in the living body of an animal or a non-animal human. For in vivo applications, the EAB sensors may include an aptamer-coated microneedle or wire as a working electrode. The microneedle or wire may be inserted through the surface of the skin such that the aptamer-coated portion contacts a biological fluid of the subcutaneous tissues. Alternatively, the EAB sensors may be placed in another bodily compartment, such as a vein. The EAB sensors can also include a counter electrode and a reference electrode. These electrodes can also be in the form of a microneedle, or a wire similarly inserted under the skin.

[0085] The EAB sensor electrodes may remain in situ for minutes, hours or even days and over that period provide clinically valuable information on the amount of analyte in the bodily fluid. Reasonable extrapolation to estimate amounts of analyte in the general circulation may be made. In this way, an EAB sensor can provide clinically relevant information on the amount of an exogenous analyte (such as a drug) or an endogenous analyte (such as a hormone) in the subject. The information may be used in the diagnosis, treatment, and/or monitoring of a disease.

[0086] The working electrode of an EAB sensor may have at least one associated counter electrode and at least one associated reference electrode. Each working electrode may have a dedicated counter electrode, however in some embodiments the counter electrode is shared amongst some or all the assembled working electrodes. Each working electrode may have a dedicated reference electrode, however in some embodiments the reference electrode is shared amongst some or all the assembled working electrodes.

[0087] In many embodiments, an EAB sensor may be voltametric, chronoamperometric, or impedimetric. In a voltametric sensor, a potential waveform is applied to the sensor interface, and the resulting current response is recorded. In chronometric approaches, a step potential is applied, and the resulting time-evolving current response is recorded. In impedimetric sensing, a sinusoidal potential waveform is applied, and the resulting sinusoidal current response is recorded.

[0088] In some embodiments, EAB sensors are of the voltametric type, with a recognition element being bound to the working electrode. Gold can be used as the probe surface for the working electrode. The recognition element can be a nucleic acid or a DNA aptamer. The recognition element has an associated redox-active species which acts as a reporter. The redox reporter can be (but is not limited to) methylene blue. Upon target (e.g., drug) binding, the recognition element undergoes a conformational change, bringing the redox reporter more proximal to the working electrode surface. This increase in proximity increases electron transfer from the redox reporter to the electrode. The increase in speed of electron transfer contributes to a change in Faradaic current that is detected by a potentiostat. EAB sensors can be incorporated into a circuit having a reference electrode. The reference electrode is the site of a known chemical reaction that has a known redox potential. For example, a reference electrode based on the silversilver chloride (Ag/AgCI) redox pair has a fixed and known potential forming the point against which the redox potential of the working electrode is measured. Also typically included in the circuit is a counter electrode which functions as a cathode or an anode to the working electrode. Because only minimal current passes through the reference electrode, the current generated is primariy attributed to the working and counter electrodes. Current is measured as a function of potential of the interrogating electrode versus the reference electrode. The difference in potential produces the current in the circuit thereby generating an output signal. The signal quantifies target binding depending on electron transfer that is ideally stoichiometrically proportional to target binding.

[0089] In some embodiments, the working electrode or any other electrode may be a wire, a needle, a microneedle, an electrode array, a microneedle array, which contact the body of a subject. Microneedles and/or microneedle arrays are preferred for transdermal applications where piercing of the skin is necessary to contact the bodily fluid.

[0090] Electrodes in accordance with many embodiments can be fabricated in a range of various shapes and geometries, although their specific geometry for transdermal applications be optimized to breach the stratum corneum for reliable skin penetration. In some embodiments, the apparatus may be configured to be urged into the skin of a subject to facilitate the electrodes breaching the stratum corneum and to penetrate through the skin layers. For non-human applications, the stratum corneum may be replaced by an analogous, or even a non-analogous layer on the surface of the subject. [0091] Ideally, feedback control will converge on the desired concentration both rapidly and accurately, without significantly overshooting it. These goals, however, reflect a trade-off. A rapid rise to the desired set point is often accompanied by a tendency to overshoot the set point, particularly when the site of measurement (and thus of concentration control) is different from the site of delivery of the controlled substance (e.g. , control over drug concentrations in the solid tissues via intravenous drug delivery). To achieve a good balance between these effects, controllers may require that their parameters be tuned based on the expected behavior of the system they are controlling. (See, e.g., Borase, R. P., et al., 2021 , International Journal of Dynamics and Control, 9(2), 818-827; the disclosure of which is incorporated by reference.) Unfortunately, accurate knowledge of the pharmacokinetics of any given subject, which often differs from subject to subject, is unlikely to be available prior to the start of delivery. This can become a problem if the tolerance for overshooting the targeted concentration is low (e.g., due to toxicity) or, as is the case for the control of concentrations in the solid tissues, transport dynamics introduce unknown delays between dosing (e.g., delivery to the blood) and the concentration rise at the site under control. In preliminary example studies of control over intracranial concentrations of the anesthetic procaine, significant overshooting could have had lethal consequences. And due to slow transport across the blood-brain-barrier, that system also faces significant delays between dosing and the appearance of the controlled substance at the site of measurement.

[0092] To account for these issues, several embodiments generated a conservative PID controller using a “population approach.” In one example, the procaine pharmacokinetics of seven animals were measured, each of which is dosed with a defined amount of the drug before monitoring the resulting rise and fall of the drug concentration in the brain without feedback control (Figure 2A). Some embodiments use simulations to design a PID controller that, given each animal’s individual pharmacokinetics, should have performed well for all seven animals. When the resulting controller was applied to an eighth, “out of training set” animal, it successfully maintains the concentration of inbrain procaine at the targeted 100 pM for about 1.66 h with a RMSD of ±18 |iM (Figure 2B). However, the controller would have failed if the pharmacokinetics of the subject under control had fallen significantly outside of the range of pharmacokinetics observed in the test population. Indeed, in the experiment the drug concentration rises to the targeted concentration more rapidly than expected given the population pharmacokinetics used to generate the controller (Figure 2C), highlighting the sometimes-difficult task of appropriately predicting the pharmacokinetics of individuals based on population data.

[0093] As a solution to the problems inherent in population-based feedback control, many embodiments implement an “adaptive control” approach that does not require the use of population data or any prior knowledge of the subject’s pharmacokinetics. In several embodiments, the feedback control algorithm actively determines the pharmacokinetics of each individual subject “on-the-fly,” while the controller is adjusting the delivery rate of the substance (i.e. , the molecule and/or ion) under control to achieve the desired time-concentration profile. Some embodiments use a procaine-detecting EAB sensor comprising a procaine-biding DNA aptamer modified with a methylene blue redox reporter and site-specifically attached to a gold electrode (Figure 1 A). The EAB sensor is inserted into the right lateral ventricles of live rats (briefly sedated using isofluorane) through a 19G cannula that can be stereotaxically implanted and cemented to the skull at least 1 week prior to the experiment. The sensor is then connected to the potentiostat via a cage-top swivel. For drug doses, some embodiments use a catheter that runs through the same swivel to a syringe pump, the pumping rate of which is modulated by the feedback controller.

[0094] The adaptive controllers in accordance with many embodiments can predict how the concentration at the site of control will respond to changes in the rate with which the substance under control is delivered into the body. To achieve this without foreknowledge (such as population-based constraints or approximations) of the subject’s specific pharmacokinetics, some embodiments assume that the future concentrations can be predicted reasonably well by a combination of recent past values of the concentration and the delivery rates (1 ). This type of model, which is called autoregressive moving average (ARMA) models, is a common way to explore relations hidden in time-series datasets. (See, e.g., Sdderstrdm, T., & Stoica, P., 2002, Circuits, Systems and Signal Processing, 21( ), 1-9; the disclosure of which is incorporated by reference.) Under a general ARMA model, the concentration at the site of control at time t, x(t), is assumed to be given by the following relation. where a_; and are the autoregressive and moving-average coefficients respectively, x(t) is the output and the u(t) is the input at time t. Such an ARMA model has an order Since the pharmacokinetics of procaine in the brain are well described by a single exponential decay, some embodiments employ an ARMA model of order (2, 1 ), that is simple enough to run optimizations but complex enough to capture previously observed single-exponential behavior.

[0095] In several embodiments, certain prior knowledge can be added to the controller such that the prior knowledge can help initialize the ARMA model, making it converge more rapidly. This could include population-based estimates of pharmacokinetics or any other information regarding the specific subject’s pharmacokinetics. In certain embodiments, the feedback controller is not initialized before the start of control. In some embodiments, the feedback controller is initialized before the start of control using population pharmacokinetic data to ensure more rapid convergence on an accurate model of the subject’s pharmacokinetics. In several embodiments, the feedback controller is initialized before the start of control using any foreknowledge or estimates of the patient’s specific pharmacokinetics to ensure more rapid convergence on an accurate model of the subject’s pharmacokinetics. [0096] To determine the coefficients of the ARMA model described by equation (1 ), the experiment starts with an initial delivery profile that is defined by the same rules for all subjects. Then, by observing how the concentration at the site of control varies in response to this initial delivery, the subject’s pharmacokinetics can be estimated. To optimize this process, the initial delivery profile is designed to provide significant information regarding the subject’s pharmacokinetic response while also rapidly moving towards to the targeted concentration. Specifically, certain embodiments employ a “pulselike” injection profile for delivery, similar to those that are optimal for estimating pharmacokinetics with high precision. (See, e.g., Erdal, M. K., et al., 2021 , 60th IEEE Conference on Decision and Control (CDC), 3072-3079; the disclosure of which is incorporated by reference.) This starts by delivering the substance at the maximum injection rate (typically defined by clinical, technical, physiological, or animal-welfare- concerns) until a pre-defined, intermediate concentration (for example, about 30 to 40% of the desired final set point) is achieved. The injection rate is then reduced by half until a second predefined, intermediate concentration (for example, about 60-70% of the final set point) is achieved. From the measurements of the subject’s response to this initial injection profile, an “individualized” pharmacokinetic model can be developed in real time. With this in hand, the adaptive controller then takes control of further delivery.

[0097] The first stage of the adaptive control approach is a system identification stage, in which the subject-specific pharmacokinetic model is updated in real time. To allow the controller to predict the future concentrations for a designed delivery rate, an ARMA (2, 1 ) model is used to explain the observed data during the initial period of delivery. However, since this model is an approximation and the real pharmacokinetics is probably more complicated (e.g., the various rate “constants” could vary with time as health-status changes), it may not provide reliable predictions arbitrarily far into the future. To circumvent this, several embodiments update the model fit after every measurement, giving the controller a chance to account for any errors that may arise from mismatch between the unknown “true” model and the approximated simple ARMA (2, 1 ) model. This approach allows the prediction of future concentrations with reasonable accuracy and precision. As an initial test of this predictive performance, certain embodiments used the data from the previous population based PID experiment (Figure 2B). Specifically, at a given time, certain embodiments implement the approach by using the recently observed concentration values and the employed injection rates (last 100-time intervals) to fit the ARMA (2, 1 ) model. This fitted model then successfully predicts future concentration values given the injection rates applied during the experiment (Figures 3A through 3C). This forms the basis of the controller since it establishes a means of predicting future concentration given the current and prior concentrations and the known prior delivery rates. The best ARMA (2, 1 ) model to the recent data can be determined by solving the nonlinear optimization problem with equality constraints. subject to x(i) = oq (i — 1) + a₂x(i — 2) + p_oiz(i — 1) + u(i — 2) where a₁₍ a₂ are autoregressive coefficients, are the moving average coefficients, x(t) is the estimated noise-free, and the y(t) is the measured, but noisy, concentration at time t . The advantage of including dynamics as equality constraints lies in computational efficiency when using second order methods to solve the relevant equation (2). The equality constraints enable this efficiency by making the Hessian of the problem much sparser. (See, e.g., Hespanha, J. P., et al., 2021 , Annual Reviews in Control, 51, 460-476; the disclosure of which is incorporated by reference.)

[0098] To achieve the targeted concentration, a model predictive controller based on the model identified in the first stage can be used. (See, e.g., Hewing, L., et al., 2020, Annual Review of Control, Robotics, and Autonomous Systems, 3(1 ), 269-296; the disclosure of which is incorporated by reference.) The idea behind model predictive control is to find the optimal delivery profile that will achieve the desired set point concentration. This “optimal control problem” can also be formulated in a manner to that of the first system identification stage. u* = argmin subject t where the ARMA model parameters (a₁, a₂> Po>^_l ) and the initial noise-free concentration (x(t), x(t - 1)) are obtained by the first identification stage formulated by equation (2). Once this optimal control problem is solved, the optimal control input iffor the next M time intervals is known. This means that u* is the best course of action to take for the identified model for the next M time intervals. However, since the identified model may not be an accurate model of the subject’s true pharmacokinetics (no matter how well it describes the recently observed concentration measurements), only the first value of u* is used at the time. At the next time step, the next newly measured noisy concentration (and the corresponding delivery rate) is added, remove the oldest data pair (concentration/delivered amount) from the consideration, and repeat the process again. This process can be summarized in the following Algorithm 1 .

Algorithm 1. Adaptive Model Predictive Algorithm

Initialization Step:

Start with r = 50,

While concentration is less than ,

Deliver at maximum injection rate

While concentration is less than C₂ ,

If we have been delivering at the same rate for T_update min, r = max(1.5 x r, 100)

Deliver at r percent of the maximum rate,

Adaptive Control Step:

While target concentration at time t is nonzero,

Estimate the approximate ARMA (2, 1 ) with a*, p* by solving (2)

Using a*, p*, find the optimal control u* by solving (3)

Apply the first value of T

Update time t = t + 1

[0099] A key set of parameters in the adaptive control algorithm is the threshold values defining when to switch control over delivery from the initial delivery profile to the adaptive controller (see C_1# C₂ at Algorithm 1 ). These values should be chosen based on (i) how rapidly the targeted concentration should be achieved and (ii) expected noise. For example, in the first attempt, for which the targeted concentration was 100 pM, the threshold is set to lower the delivery rate, , to 35 pM and the threshold to start the adaptive controller, C₂, to 60 pM (Figure 4). These thresholds are lower than the targeted concentration to prevent build-up of the substance to a level at which the adaptive controller cannot avoid overshooting. If the concentration may rise much more slowly (or rapidly), C₂ can be set higher (or lower). After this initialization, the adaptive algorithm takes over and computes the optimal delivery rate at each time step and continues updating until the experiment ends. [00100] The adaptive feedback control approach supports the rapid, accurate control over in-brain procaine concentrations (Figures 5A through 5D). For example, using this approach, some embodiments attempt to hold in-brain procaine concentrations at about 100 pM for about 1 h. The adaptive controller achieved the target concentration about 1 h later. It then held this concentration with a RMSD of 10.2 pM (Figure 5A). This time-to- set point and control precision represent about a 2-fold improvement over the performance of earlier, population-pharmacokinetics-with-a-PID-controller efforts. This improvement in performance is achieved without employing any prior knowledge of the drug’s pharmacokinetics at the population level, much less at the level of the individual subject.

[00101] The ability to perform control experiments with no prior knowledge of the subject’s pharmacokinetics allows successful feedback control experiments with a range of concentration time courses without having to build population-based models first. Some embodiments repeat the same experiment but target a fixed set point of about 200 pM (Figure 5B). The adaptive controller takes over at 12.4 min after the start of delivery and achieves its target concentration about 9.7 min later. It then maintains the targeted concentration (RMSD of 11.3 pM ) until control is released after about 1 h. Some embodiments use the algorithm to respond to a changing set concentration (i.e. , a stairstep profile). In the first demonstration, the concentration is raised to about 100 pM (in 32 min), holds it at this level (RMSD 4.9 pM) for 31 min, and then reduces the set point to about 50 pM, holding this (RMSD 4.1 pM) for an additional 38 min (Figure 5C). In the last experiment, the targeted concentration is stepped in the opposite direction (Figure 5D). A concentration of about 50 pM is targeted, which is achieved after 33 min and held (RMSD 6.0 pM) for 42 min. The set point concentration is then raised to about 100 pM, achieving this after an additional 16 min, and holds it (RMSD 11 .3 pM) for about 1 h before ending the experiment.

[00102] The real-time concentration information provided by in vivo EAB sensors provides an opportunity to implement advanced control techniques to achieve high precision delivery of drugs and other biologically important molecules and ions. Subject- to-subject and time-varying intrasubject pharmacokinetic variability makes designing a controller, that will achieve targeted concentration time profiles challenging, particularly when the site of control is in a different compartment from the site of delivery. Many embodiments show that an adaptive controller can achieve targeted concentration profiles without requiring any foreknowledge of the subject’s pharmacokinetics or of the population pharmacokinetics of the substance under control.

EXAMPLES

[00103] Example 1 : An adaptive feedback control system, comprising: an electrochemical aptamer based (EAB) sensor, comprising: a working electrode, wherein the working electrode comprises a recognition element modified with a redox molecule; wherein the recognition element specifically binds to a substance and is configured to undergo a conformational change upon binding with the molecule or ion; wherein a current readout from the EAB sensor at a site of control reflects a concentration of the substance; a delivery means; wherein the delivery means is configured to deliver the substance to a site of delivery; and a feedback controller; wherein the feedback controller is configured to modulate a delivery rate of the delivery means; wherein the feedback controller predicts pharmacokinetics of the substance at the site of control in response to the delivery rate and the concentration of the substance such that the feedback controller adjusts the delivery rate in real time to achieve a predefined concentration of the substance at the site of control.

[00104] Example 2: The system of example 1 , wherein the site of control and the site of delivery are different.

[00105] Example 3: The system of example 1 or 2, wherein the site of control is a body part or a region of a subject.

[00106] Example 4: The system of example 1 , or 2, or 3, wherein the site of control is a site of drug action. [00107] Example 5: The system of any one of examples 1 to 4, wherein the site of control is the brain of a subject, and the site of delivery is selected from the group consisting of: a vein, an artery, a muscle, an intraperitoneal space, and a subcutaneous space.

[00108] Example 6: The system of any one of examples 1 to 5, wherein the substance comprises a plurality of molecules, or a plurality of ions.

[00109] Example 7: The system of any one of examples 1 to 6, wherein the feedback controller implements an initial delivery profile by injecting the substance at a feasible injection rate until the delivery means achieves a predefined intermediate concentration of the substance; after the initial injection, the feedback controller reduces the injection rate by a pre-defined amount until the delivery achieves a second predefined, intermediate concentration; wherein the injection rate is clinically feasible, or technically feasible, or physiologically feasible, or animal-welfare appropriate.

[00110] Example 8: The system of any one of examples 1 to 7, wherein the pre-defined amount is by half.

[00111] Example 9: The system of any one of examples 1 to 8, wherein the predefined intermediate concentration is between 30% and 40% of a predefined desired concentration of the substance at the site of control; wherein the second predefined intermediate concentration is between 60% and 70% of the predefined desired concentration of the substance at the site of control.

[00112] Example 10: The system of any one of examples 1 to 9, wherein after the feedback controller reaches a threshold value, the feedback controller switches from the initial injection profile to an optimal injection profile.

[00113] Example 11 : The system of any one of examples 1 to 10, wherein the optimal injection profile uses an optimal control input calculated by solving a mathematical equation at a first time step; at a time step after the first time step, the feedback controller uses a measured concentration and a corresponding injection rate to replace the optimal control input such that the feedback controller updates and adjusts the delivery rate to an optimal delivery rate at each time step to achieve the predefined concentration at the site of control.

[00114] Example 12: The system of any one of examples 1 to 11 , wherein the feedback controller does not use population data-based pharmacokinetics of the substance.

[00115] Example 13: The system of any one of examples 1 to 12, wherein the feedback controller does not use prior measurements of pharmacokinetics of the substance.

[00116] Example 14: The system of any one of examples 1 to 13, wherein the feedback controller is not initialized before a start of control.

[00117] Example 15: The system of any one of examples 1 to 14, wherein the site of control is a body part of a subject and the feedback controller is initialized before a start of control using population pharmacokinetic data to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

[00118] Example 16: The system of any one of examples 1 to 15, wherein the site of control is a body part of a subject and the feedback controller is initialized before a start of control using foreknowledge or an estimate of a subject’s specific pharmacokinetics to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

[00119] Example 17: The system of any one of examples 1 to 16, wherein the delivery means is a catheter or a syringe.

[00120] Example 18: The system of any one of examples 1 to 17, wherein the recognition element is a nucleic acid.

[00121] Example 19: The system of any one of examples 1 to 18, wherein the recognition element is a DNA aptamer.

[00122] Example 20: The system of any one of examples 1 to 19, wherein the working electrode is a wire, a rod, a planar electrode, a needle, or a microneedle.

[00123] Example 21 : The system of any one of examples 1 to 20, wherein the conformational change alters an electron transfer rate between the redox molecule and a surface of the working electrode. [00124] Example 22: The system of any one of examples 1 to 21 , wherein the substance is a drug molecule selected from the group consisting of: a small molecule drug molecule, an amino acid-based drug molecule, a protein-based drug molecule, a lipid-based drug, a polysaccharide-based drug and a nuclear acid-based drug molecule. [00125] Example 23: The system of any one of examples 1 to 22, wherein the substance is a non-drug molecule selected from the group consisting of: a hormone, an enzyme, an antibody, a cytokine, and a metabolite.

[00126] Example 24: The system of any one of examples 1 to 23, wherein the substance is an ion selected from the group consisting of sodium, potassium, lithium, calcium, magnesium, and chloride.

[00127] Example 25: The system of any one of examples 1 to 24, wherein the substance is procaine, the recognition element is a procaine-binding DNA aptamer, the redox molecule is methylene blue, the site of delivery is a peripheral tissue, and the site of control is the brain.

[00128] Example 26: The system of any one of examples 1 to 25, wherein the EAB sensor further comprises a counter and reference electrode pair, or a counter electrode and a separate reference electrode.

[00129] Example 27: A method for adaptive feedback control, comprising: delivering a substance to a cite of delivery; measuring a concentration of the substance using at least one EAB sensor; wherein the at least one EAB sensor comprises a working electrode comprising a recognition element modified with a redox molecule; wherein the recognition element specifically binds to the substance and is configured to undergo a conformational change upon binding; wherein a current readout from the EAB sensor at a site of control reflects a concentration of the substance; and controlling a delivery rate via a feedback controller, wherein the feedback controller predicts pharmacokinetics of the substance at the site of control in response to the delivery rate and the concentration of the substance such that the feedback controller adjusts the delivery rate in real time to achieve a predefined concentration of the substance at the site of control.

[00130] Example 28: The method of example 27, wherein the site of control and the site of delivery are different.

[00131] Example 29: The method of example 27 or 28, wherein the site of control is a body part of a subject.

[00132] Example 30: The system of example 27, or 28, or 29, wherein the site of control is a site of drug action.

[00133] Example 31 : The system of any one of examples 27 to 30, wherein the site of control is the brain of a subject, and the site of delivery is selected from the group consisting of: a vein, an artery, a muscle, an intraperitoneal space, a subcutaneous space, and a peripheral tissue of the subject.

[00134] Example 32: The system of any one of examples 27 to 31 , wherein the feedback controller implements an initial injection profile by delivering the substance at a feasible injection rate until the delivery achieves a predefined intermediate concentration; after the initial injection, the feedback controller reduces the injection rate until the delivery achieves a second predefined, intermediate concentration; wherein the injection rate is clinically feasible, or technically feasible, or animal-welfare appropriate.

[00135] Example 33: The system of any one of examples 27 to 32, wherein the predefined intermediate concentration is between 30% and 40% of a predefined desired concentration at the site of control; wherein the second predefined intermediate concentration is between 60% and 70% of the predefined desired concentration at the site of control.

[00136] Example 34: The system of any one of examples 27 to 33, wherein after the feedback controller reaches a threshold value, the feedback controller switches from the initial injection profile to an optimal injection profile.

[00137] Example 35: The system of any one of examples 27 to 34, wherein the optimal injection profile uses an optimal control input calculated by solving a mathematical equation at a first time step; at a time step after the first time step, the feedback controller uses a measured concentration and a corresponding injection rate to replace the optimal control input such that the feedback controller updates and adjusts the delivery rate to an optimal delivery rate at each time step to achieve predefined concentration of the substance at the site of control.

[00138] Example 36: The system of any one of examples 27 to 35, wherein the feedback controller does not use population data-based pharmacokinetics of the substance.

[00139] Example 37: The system of any one of examples 27 to 36, wherein the feedback controller is not initialized before a start of control.

[00140] Example 38: The system of any one of examples 27 to 37, wherein the site of control is a body part of a subject and the feedback controller is initialized before a start of control using population pharmacokinetic data to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

[00141] Example 39: The system of any one of examples 27 to 38, wherein the site of control is a body part of a subject and the feedback controller is initialized before a start of control using foreknowledge or estimates of a subject’s specific pharmacokinetics to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

[00142] Example 40: The system of any one of examples 27 to 39, wherein the feedback controller does not use prior measurements of pharmacokinetics of the substance.

[00143] Example 41 : The system of any one of examples 27 to 40, wherein the delivery is via a catheter or a syringe.

[00144] Example 42: The system of any one of examples 27 to 41 , wherein the recognition element is a nucleic acid.

[00145] Example 43: The system of any one of examples 27 to 42, wherein the recognition element is a DNA aptamer.

[00146] Example 44: The system of any one of examples 27 to 43, wherein the working electrode is a wire, a rod, a planar electrode, a needle, or a microneedle. [00147] Example 45: The system of any one of examples 27 to 44, wherein the conformational change alters an electron transfer rate between the redox molecule and a surface of the working electrode.

[00148] Example 46: The system of any one of examples 27 to 45, wherein the at least one EAB sensor further comprises a counter and reference electrode pair, or a counter electrode and a separate reference electrode.

[00149] Example 47: The system of any one of examples 27 to 46, wherein the substance comprises a plurality of molecules, or a plurality of ions.

[00150] Example 48: The system of any one of examples 27 to 47, wherein the substance is a drug molecule selected from the group consisting of: a small molecule drug molecule, an amino acid-based drug molecule, a protein-based drug molecule, a lipid-based drug molecule, a polysaccharide-based drug molecule and a nuclear acidbased drug molecule.

[00151] Example 49: The system of any one of examples 27 to 48, wherein the substance is a non-drug molecule selected from the group consisting of: a hormone, an enzyme, an antibody, a cytokine, and a metabolite.

[00152] Example 50: The system of any one of examples 27 to 49, wherein the substance is an ion selected from the group consisting of sodium, potassium, lithium, calcium, magnesium, and chloride.

[00153] Example 51 : The system of any one of examples 27 to 50, wherein the substance is procaine, the recognition element is a procaine-binding DNA aptamer, the redox molecule is methylene blue, the site of delivery is a peripheral tissue, and the site of control is the brain.

DOCTRINE OF EQUIVALENTS

[00154] As can be inferred from the above discussion, the above-mentioned concepts can be implemented in a variety of arrangements in accordance with embodiments of the invention. Accordingly, although the present invention has been described in certain specific aspects, many additional modifications and variations would be apparent to those skilled in the art. It is therefore to be understood that the present invention may be practiced otherwise than specifically described. Thus, embodiments of the present invention should be considered in all respects as illustrative and not restrictive.

[00155] As used herein, the singular terms “a,” “an,” and “the,” may include plural referents unless the context clearly dictates otherwise. Reference to an object in the singular is not intended to mean "one and only one" unless explicitly so stated, but rather “one or more.”

[00156] As used herein, the terms “approximately,” and “about” are used to describe and account for small variations. When used in conjunction with an event or circumstance, the terms can refer to instances in which the event or circumstance occurs precisely as well as instances in which the event or circumstance occurs to a close approximation. When used in conjunction with a numerical value, the terms can refer to a range of variation of less than or equal to ± 10% of that numerical value, such as less than or equal to ±5%, less than or equal to ±4%, less than or equal to ±3%, less than or equal to ±2%, less than or equal to ±1 %, less than or equal to ±0.5%, less than or equal to ±0.1 %, or less than or equal to ±0.05%.

[00157] Additionally, amounts, ratios, and other numerical values may sometimes be presented herein in a range format. It is to be understood that such range format is used for convenience and brevity and should be understood flexibly to include numerical values explicitly specified as limits of a range, but also to include all individual numerical values or sub-ranges encompassed within that range as if each numerical value and sub-range is explicitly specified. For example, a ratio in the range of about 1 to about 200 should be understood to include the explicitly recited limits of about 1 and about 200, but also to include individual ratios such as about 2, about 3, and about 4, and sub-ranges such as about 10 to about 50, about 20 to about 100, and so forth.

Claims

CLAIMS:

1 . An adaptive feedback control system, comprising: an electrochemical aptamer based (EAB) sensor, comprising: a working electrode, wherein the working electrode comprises a recognition element modified with a redox molecule; wherein the recognition element specifically binds to a substance and is configured to undergo a conformational change upon binding with the molecule or ion; wherein a current readout from the EAB sensor at a site of control reflects a concentration of the substance; a delivery means; wherein the delivery means is configured to deliver the substance to a site of delivery; and a feedback controller; wherein the feedback controller is configured to modulate a delivery rate of the delivery means; wherein the feedback controller predicts pharmacokinetics of the substance at the site of control in response to the delivery rate and the concentration of the substance such that the feedback controller adjusts the delivery rate in real time to achieve a predefined concentration of the substance at the site of control.

2. The system of claim 1 , wherein the site of control and the site of delivery are different.

3. The system of claim 1 , wherein the site of control is a body part or a region of a subject.

4. The system of claim 1 , wherein the site of control is a site of drug action.

5. The system of claim 1 , wherein the site of control is the brain of a subject, and the site of delivery is selected from the group consisting of: a vein, an artery, a muscle, an intraperitoneal space, and a subcutaneous space.

6. The system of claim 1 , wherein the substance comprises a plurality of molecules, or a plurality of ions.

7. The system of claim 1 , wherein the feedback controller implements an initial delivery profile by injecting the substance at a feasible injection rate until the delivery means achieves a predefined intermediate concentration of the substance; after the initial injection, the feedback controller reduces the injection rate by a pre-defined amount until the delivery achieves a second predefined, intermediate concentration; wherein the injection rate is clinically feasible, or technically feasible, or physiologically feasible, or animal-welfare appropriate.

8. The system of claim 7, wherein the pre-defined amount is by half.

9. The system of claim 7, wherein the predefined intermediate concentration is between 30% and 40% of a predefined desired concentration of the substance at the site of control; wherein the second predefined intermediate concentration is between 60% and 70% of the predefined desired concentration of the substance at the site of control.

10. The system of claim 7, wherein after the feedback controller reaches a threshold value, the feedback controller switches from the initial injection profile to an optimal injection profile.

11. The system of claim 10, wherein the optimal injection profile uses an optimal control input calculated by solving a mathematical equation at a first time step; at a time step after the first time step, the feedback controller uses a measured concentration and a corresponding injection rate to replace the optimal control input such that the feedback controller updates and adjusts the delivery rate to an optimal delivery rate at each time step to achieve the predefined concentration at the site of control.

12. The system of claim 1 , wherein the feedback controller does not use population data- based pharmacokinetics of the substance.

13. The system of claim 1 , wherein the feedback controller does not use prior measurements of pharmacokinetics of the substance.

14. The system of claim 1 , wherein the feedback controller is not initialized before a start of control.

15. The system of claim 1 , wherein the site of control is a body part of a subject and the feedback controller is initialized before a start of control using population pharmacokinetic data to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

16. The system of claim 1 , wherein the site of control is a body part of a subject and the feedback controller is initialized before a start of control using foreknowledge or an estimate of a subject’s specific pharmacokinetics to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

17. The system of claim 1 , wherein the delivery means is a catheter or a syringe.

18. The system of claim 1 , wherein the recognition element is a nucleic acid.

19. The system of claim 1 , wherein the recognition element is a DNA aptamer.

20. The system of claim 1 , wherein the working electrode is a wire, a rod, a planar electrode, a needle, or a microneedle.

21. The system of claim 1 , wherein the conformational change alters an electron transfer rate between the redox molecule and a surface of the working electrode.

22. The system of claim 1 , wherein the substance is a drug molecule selected from the group consisting of: a small molecule drug molecule, an amino acid-based drug molecule, a protein-based drug molecule, a lipid-based drug, a polysaccharide-based drug and a nuclear acid-based drug molecule.

23. The system of claim 1 , wherein the substance is a non-drug molecule selected from the group consisting of: a hormone, an enzyme, an antibody, a cytokine, and a metabolite.

24. The system of claim 1 , wherein the substance is an ion selected from the group consisting of sodium, potassium, lithium, calcium, magnesium, and chloride.

25. The system of claim 1 , wherein the substance is procaine, the recognition element is a procaine-binding DNA aptamer, the redox molecule is methylene blue, the site of delivery is a peripheral tissue, and the site of control is the brain.

26. The system of claim 1 , wherein the EAB sensor further comprises a counter and reference electrode pair, or a counter electrode and a separate reference electrode.

27. A method for adaptive feedback control, comprising: delivering a substance to a cite of delivery; measuring a concentration of the substance using at least one EAB sensor; wherein the at least one EAB sensor comprises a working electrode comprising a recognition element modified with a redox molecule; wherein the recognition element specifically binds to the substance and is configured to undergo a conformational change upon binding; wherein a current readout from the EAB sensor at a site of control reflects a concentration of the substance; and controlling a delivery rate via a feedback controller; wherein the feedback controller predicts pharmacokinetics of the substance at the site of control in response to the delivery rate and the concentration of the substance such that the feedback controller adjusts the delivery rate in real time to achieve a predefined concentration of the substance at the site of control.

28. The method of claim 27, wherein the site of control and the site of delivery are different.

29. The method of claim 27, wherein the site of control is a body part of a subject.

30. The method of claim 27, wherein the site of control is a site of drug action.

31. The method of claim 27, wherein the site of control is the brain of a subject, and the site of delivery is selected from the group consisting of: a vein, an artery, a muscle, an intraperitoneal space, a subcutaneous space, and a peripheral tissue of the subject.

32. The method of claim 27, wherein the feedback controller implements an initial injection profile by delivering the substance at a feasible injection rate until the delivery achieves a predefined intermediate concentration; after the initial injection, the feedback controller reduces the injection rate until the delivery achieves a second predefined, intermediate concentration; wherein the injection rate is clinically feasible, or technically feasible, or animal-welfare appropriate.

33. The method of claim 32, wherein the predefined intermediate concentration is between 30% and 40% of a predefined desired concentration at the site of control; wherein the second predefined intermediate concentration is between 60% and 70% of the predefined desired concentration at the site of control.

34. The method of claim 32, wherein after the feedback controller reaches a threshold value, the feedback controller switches from the initial injection profile to an optimal injection profile.

35. The method of claim 34, wherein the optimal injection profile uses an optimal control input calculated by solving a mathematical equation at a first time step; at a time step after the first time step, the feedback controller uses a measured concentration and a corresponding injection rate to replace the optimal control input such that the feedback controller updates and adjusts the delivery rate to an optimal delivery rate at each time step to achieve predefined concentration of the substance at the site of control.

36. The method of claim 27, wherein the feedback controller does not use population data- based pharmacokinetics of the substance.

37. The method of claim 27, wherein the feedback controller is not initialized before a start of control.

38. The method of claim 27, wherein the site of control is a body part of a subject and the feedback controller is initialized before a start of control using population pharmacokinetic data to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

39. The method of claim 27, wherein the site of control is a body part of a subject and the feedback controller is initialized before a start of control using foreknowledge or estimates of a subject’s specific pharmacokinetics to achieve a more rapid convergence on an accurate model of the subject’s pharmacokinetics.

40. The method of claim 27, wherein the feedback controller does not use prior measurements of pharmacokinetics of the substance.

41 . The method of claim 27, wherein the delivery is via a catheter or a syringe.

42. The method of claim 27, wherein the recognition element is a nucleic acid.

43. The method of claim 27, wherein the recognition element is a DNA aptamer.

44. The method of claim 27, wherein the working electrode is a wire, a rod, a planar electrode, a needle, or a microneedle.

45. The method of claim 27, wherein the conformational change alters an electron transfer rate between the redox molecule and a surface of the working electrode.

46. The method of claim 27, wherein the at least one EAB sensor further comprises a counter and reference electrode pair, or a counter electrode and a separate reference electrode.

47. The method of claim 27, wherein the substance comprises a plurality of molecules, or a plurality of ions.

48. The method of claim 27, wherein the substance is a drug molecule selected from the group consisting of: a small molecule drug molecule, an amino acid-based drug molecule, a protein-based drug molecule, a lipid-based drug molecule, a polysaccharide-based drug molecule and a nuclear acid-based drug molecule.

49. The method of claim 27, wherein the substance is a non-drug molecule selected from the group consisting of: a hormone, an enzyme, an antibody, a cytokine, and a metabolite.

50. The method of claim 27, wherein the substance is an ion selected from the group consisting of sodium, potassium, lithium, calcium, magnesium, and chloride.

51. The method of claim 27, wherein the substance is procaine, the recognition element is a procaine-binding DNA aptamer, the redox molecule is methylene blue, the site of delivery is a peripheral tissue, and the site of control is the brain.