Method and sensor

Abstract

A sensor 100 for extracting and detecting an analyte (e.g. glucose, hormones, toxins, biomarkers) from a subject comprises at least one co-monitoring sensor pixel 110, a substrate 102 for positioning

Description

Field of the Invention The technology relates to methods and devices for estimating a concentration of a substance in a subject. This includes the detection and monitoring of substances such as glucose, hormones, medications, markers, toxins, and specific molecules or compounds present in a subject’s body, e.g. in a subdermal region of the subject’s body. Background The monitoring and measurement of subdermal analytes, such as glucose, hormones, medication concentrations, and other biomarkers, play a crucial role in the diagnosis, treatment, and management of various medical conditions. Traditional methods for determining subdermal analyte levels often involve invasive procedures, such as blood draws or tissue biopsies, which can be painful, timeconsuming, and may pose risks of infection or other complications. Implantable subdermal sensors may also provide similar risks. Non-invasive techniques for measuring subdermal analytes have been developed, including transdermal analyte monitoring systems that rely on the passive diffusion of analytes through the skin. However, these methods often suffer from low sensitivity and specificity, as well as slow response times, due to the limited permeability of the skin and the complex nature of the skin's barrier function. Another approach to non-invasive subdermal analyte monitoring involves the use of iontophoresis, a technique that applies a small electrical current to the skin to enhance the transport of charged molecules across the skin barrier. While iontophoresis has shown promise in improving the sensitivity and specificity of transdermal analyte measurements, it still faces challenges in accurately quantifying the amount of analyte transported across the skin and in accounting for variations in skin properties and analyte transport mechanisms among different individuals. Moreover, the prior art methods often rely on the assumption that the analyte transport across the skin is uniform and consistent, which may not always be the case due to the presence of skin appendages, such as hair follicles (and associated sebaceous glands) or sweat glands that can create preferential pathways for analyte transport. This can lead to inaccuracies in the determination of subdermal analyte levels and may limit the clinical utility of these non-invasive monitoring techniques. Therefore, there is a need for an improved method for determining analyte levels in a patient that overcomes the limitations of the prior art, provides accurate and reliable measurements, and can account for the complex and heterogeneous nature of the skin and its transport mechanisms. The present invention has been devised in light of the above considerations. Summary of the Invention According to a first aspect of the invention, there is provided a computer-implemented method for estimating a concentration (e.g. a subdermal concentration) of a target substance in a subject, the method comprising: receiving a primary (first, target) electric signal corresponding to an amount of the target substance that is extracted into a reservoir of a sensor through a set of dermal channels; determining a number of dermal channels N in the set of dermal channels; determining a flux of the target substance jtarget through a selected number of dermal channels n, based on the primary electric signal and the determined number of dermal channels N; obtaining a reference flux jreference of a reference substance through the selected number of dermal channels n at a reference concentration Creference_subject of the reference substance; and estimating a concentration Ctarget_subject of the target substance in the subject based on the flux of the target substance jtarget, reference flux jreference, and reference concentration Creference_subject. The primary electric signal may comprise a target current Itarget or a target voltage Utarget that corresponds (e.g. the value of current (or voltage) of the current signal (or voltage signal) corresponds) to the amount of the target substance that is extracted into the reservoir through the set of dermal channels. For example, in embodiments utilising chronoamperometry to detect the target substance, the primary electric signal may comprise a target current Itargef 'n embodiments utilising potentiometry to detect the target substance, the primary electric signal may comprise a target voltage U target• By using an evaluated flux of the target substance jtarget in combination with the reference flux of a reference substance jreference, where both fluxes are determined through the same number of dermal channels n, and where the reference flux corresponds to a reference concentration Creference subject, the concentration of the target substance Ctarget subject in the subject can be estimated without requiring the subject to perform a separate invasive measurement e.g. to calibrate a relationship between the primary electric signal (e.g. a target voltage Utarget or target current Itarget) and the concentration of the target substance in the body. The concentration of target substance can therefore be estimated accurately and reliably, with reduced reliance on the user to perform invasive measurements (e.g. via invasive finger stick tests for calibration). The computer-implemented method may be used to perform continuous monitoring of the target substance. The computer-implemented method may be performed by a processing unit. The processing unit may have at least one processor and at least one memory including computer program code, wherein the computer program code is configured to, with the at least one processor, cause the processing unit to perform any of the steps of the computer-implemented method described herein. The processing unit may be attached to the sensor or remote from the sensor. The processing unit may be configured for communication with the sensor, e.g. for wired or wireless communication therewith. The processing unit may be configured to receive the primary electric signal from the sensor. The reference concentration Creference_subject and estimated target concentration Ctargetsubject may relate to substantially the same region in the subject’s body, e.g. both relating to subdermal concentrations in the subject’s body. As used herein, the term “flux” may refer to a mass of substance that moves through an area per unit time. For example, the area may correspond to the cross-sectional area of one or more dermal channels. The exact areas of the dermal channel may be unknown. However, by comparing a reference flux and target flux that both relate to the same number n of dermal channels that the substances are extracted through, the impact of geometrical characteristics of the dermal channels on flux may be reduced or cancelled out. As used herein, the term “target substance” may refer to any substance (analyte) within a subject. The target substance may be in a region of the subject below a principal barrier layer of skin (e.g. below the stratum corneum, which is the outermost layer of skin). The region may be the epidermis or dermis (or both), or even in appendages (skin appendages) that are located in these areas, and within the interstitial fluid which bathes these regions. Collectively, one or more of these regions may be referred to as a “subdermal region” herein. The target substance may be selected from the list comprising: glucose, a hormone, a medication / drug, a marker or toxin, or any other molecule or compound in the subject. Likewise, the term “reference substance” may also refer to any such substance. In some embodiments, the target substance and the reference substance may be the same. In such embodiments, the reference concentration Creference subject may be an expected concentration of the target substance at a particular (reference) flux, jreference of the target substance. The measured flux ]target of the target substance can thus be compared to the reference concentration and reference flux to estimate the actual concentration of the target substance in the subject. In other embodiments, the target substance and the reference substance may be different. In such embodiments, the reference concentration Creference subject may be an expected / assumed concentration of the reference substance in the subject. The measured flux jtarget of the target substance can thus be compared to the (assumed) reference concentration Creferencesubject and the measured reference flux jreference to estimate the actual concentration of the target substance in the subject. The target substance may also be referred to as a “primary substance” or “first substance”. Terms such as “target electric signal” or “target flux”, and similar, are used herein as labels to refer to electric signals, fluxes, etc that relate to the target substance (rather than the reference substance). These labels may equally be replaced e.g. with terms such as “first electric signal / flux” or “primary electric signal / flux.” The reference substance may also be referred to as a “comparison substance” or a “secondary substance”. Terms such as “reference electric signal” and “reference flux”, etc are used herein as labels to refer to electric signals, fluxes, etc that relate to the reference substance (rather than the target substance). These labels may equally be replaced e.g. with terms such as “comparison / secondary electric signal” or “comparison / secondary flux”. As used herein, the term “dermal channel” (or “transdermal channel”) may refer to any pathway suitable for extracting (e.g. transdermally iontophoretically extracting) the target substance into a reservoir of the sensor. As used herein, the term “transdermal” refers to movement through one or more layers of the skin. The dermal channel(s) may be configured to extract the target substance from a region in the subject that is below the stratum corneum, e.g. from a subdermal region of the subject. The dermal channel(s) may extend through one or more layers of the skin, with an exit at the skin surface (e.g. at the stratum corneum). The dermal channel(s) may extend through the epidermis and optionally some or all of the dermis. In some embodiments, the dermal channel(s) may be pre-existing and naturally formed (e.g. naturally occurring) in the subject, e.g. comprising a skin appendage(s) such as a hair follicle(s) (optionally including associated sebaceous gland(s)) or sweat gland(s) . This allows for true non-invasive sensing, since it is not required for any channel (e.g. needle) to be inserted within the skin itself. Alternatively, or in combination, in some embodiments, the dermal channel(s) may be formed by the sensor itself, e.g. in the form of one or more micro-needles or capillary tubes which are insertable into the skin. Although this is more invasive than using naturally-occurring dermal channels (e.g. skin appendages), these embodiments still provide advantages in that they can be used for long-term monitoring without requiring separate invasive calibration (e.g. finger-stick calibration) which is otherwise common in existing in vivo sensors. As used herein, the term “transdermal iontophoretic extraction” may refer to a process of applying a current to skin to cause substances to be extracted through the skin by electromigration or electroosmosis. In response, neutral substances (e.g. neutral molecules such as urea or glucose) are primarily extracted via electroosmosis (which occurs primarily through dermal channels), whereas charged substances (e.g. charged molecules such as ketones or lactate, or ions such as Na+) are primarily extracted via electromigration (which can occur across all regions of skin, not limited to dermal channels). Optionally, the target substance is neutral (e.g. non-charged) and the reference substance is either neutral or charged (e.g. non-neutral). As noted above, the target (first, primary) electric signal corresponds to (is indicative of, correlated to, proportional to) an amount (e.g. mass, concentration) of the target substance that is extracted into the reservoir through the set of dermal channels. For example, this can be represented by the relationship Itarget = otarget ■ Ctarget reservoir, where otarget is the (known) sensitivity of the sensor to the target substance, and Ctarget reservoir is the concentration of the target substance in the (known) volume V of the reservoir. It will be noted that the mass m of the substance in the reservoir is provided by m. Ctarget_reservoir X V. In embodiments utilising potentiometry rather than chronoamperometry to detect the target substance, a similar relationship can be represented as Utarget = otarget ■ Ctargetreservoir. The sensitivity ^target may be represented in units of current or voltage to suit the relevant equation (e.g. for use in chronoamperometry or potentiometry). Estimating the concentration Ctargetsubject of the target substance in the subject may utilise a proportionality constant to compare the target concentration to the reference concentration. Estimating the concentration may comprise an evaluation of Ctargetsubject = .}target x Creference-subject x where K is a predetermined proportionality constant between the reference substance and target substance. In embodiments where the target substance is the same as the reference substance, K = 1. In embodiments where the target substance is different from the reference substance, K may be different from 1. The value of K can be determined experimentally for any given combination of reference / target substances using a set of universally valid experiments. For example, K may be determined experimentally for a given combination of reference and target substances by measuring a concentration of the reference substance and a concentration of the target substance (e.g. using any known invasive method), using those concentrations to determine the flux of the reference substance and the flux of the target substance, and then evaluating K = Ctarget subject x ireference|n manner proportionality constant K can be determined for any ^reference subject J target given combination of reference and target substances and can subsequently be applied for estimating concentrations according to the methods described herein. The value of K may be stored e.g. in a memory of a processing unit configured to execute the computer-implemented method. Optionally, embodiment methods may include a step of obtaining the proportionality constant K from the memory (e.g. using a look-up table). As noted above, the method includes determining a number of dermal channels N in the set of dermal channels. In embodiments where the dermal channels form part of the sensor itself (e.g. as an array of channels such as micro-needles), this determination can be straightforward, based on a pre-defined number of channels that is known to be present in the sensor. In other embodiments which utilise pre-formed or naturally-occurring dermal channels within the subject, the method may include additional steps to determine the number of dermal channels that are used to extract the target substance into the reservoir. Optionally, determining the number of channels N comprises: receiving a plurality of (primary / target) electric signals each corresponding to an amount of a substance (e.g. the target substance) that is extracted into a respective reservoir of a plurality of reservoirs of the sensor; comparing the plurality of electric signals (e.g. comparing electric signals of the plurality of electric signals with other electric signals of the plurality of electric signals) to quantise a value corresponding to (indicating, proportional to) an amount of the substance per dermal channel (e.g. as a value for current, voltage, or mass per dermal channel); and determining the number of channels for a given reservoir based on the electric signal corresponding to (e.g. obtained from) said reservoir and the quantised value. Each electric signal may comprise a current or a voltage that corresponds to the amount of the substance that is extracted into the respective reservoir. By obtaining a plurality of electric signals, which each correspond to an amount of the same substance (e.g. the target substance) in a respective reservoir, the signal value (e.g. current, voltage) for each reservoir will be proportional to the number of dermal channels from which the reservoir extracts the substance. By statistically evaluating the different signals, this can allow a value corresponding to an amount of substance per dermal channel to be quantised (e.g. as a smallest common factor between all signal values). Since the density of skin appendages (e.g. hair follicles, sweat glands, etc), can differ from person-to-person (e.g. across different races, sexes, ages, etc), this method can be used to reliably determine the number of dermal channels for each subject. Optionally, each electric signal of the plurality of electric signals relates to an amount of target substance that is extracted into a different reservoir. Optionally, the plurality of electric signals comprises the same number of electric signals as the number of reservoirs. In other embodiments, the plurality of electric signals may relate to an amount of a reference substance that is extracted into different reservoirs, e.g. if the sensor is configured to co-monitortwo or more substances (as will be described further herein). The number of dermal channels for a given reservoir may be determined by dividing the current, voltage, or mass obtained from said reservoir by a respective quantised current, voltage, or mass per dermal channel. The accuracy and reliability of this method can be further improved based on the structure of the sensor device itself. For example, using a larger number of reservoirs (i.e. obtaining a larger plurality of electric signals) can help to improve the reliability for quantising the amount of substance per dermal channel, by reducing the probability of multiple reservoirs covering (extracting from) the same number of dermal channels. Optionally, the method may utilise an even number of reservoirs used to detect the target substance (e.g. to detect only the target substance). For example, optionally, the method may utilise 4 or more of said reservoirs (e.g. corresponding to 4 or more measurements for the target substance), optionally 6 or more reservoirs, optionally 8 or more, optionally 10 or more, optionally 12 or more, optionally 14 or more, optionally 16 or more, optionally 18 or more, optionally 20 or more, optionally 30 or more, optionally 40 or more, optionally 50 or more, optionally 60 or more. The reservoirs may be formed in an array of the sensor, e.g. optionally as a 2x2 array, optionally 2x3, optionally 2x4, optionally 2x5, optionally 3x4, optionally 4x4, optionally 3x6, optionally 4x5, optionally 5x5, optionally 6x6, optionally 7x7, optionally 8x8. A method utilising at least 8 measurements and up to 16 measurements (e.g. with a 4x2, 4x3, or 4x4 sensor array) may be particularly advantageous to offer redundancy and / or permit co-monitoring of a second (reference) substance, while also balancing this with a relatively small and inexpensive device compared to devices having larger array sizes. Optionally, one or more (e.g. a majority or all) of the reservoirs may be substantially cylindrical. Optionally, one or more (e.g. a majority or all) of the reservoirs may have a surface area of at least 2 mm2, optionally up to 15 mm2, e.g. optionally up to 10 mm2. These dimensions help improve the likelihood that the reservoirs will cover between 1 and 5 dermal channels, assuming an average dermal channel density (e.g. average hair follicle density) of 50 channels per cm2, with a 10 mm2 reservoir on average covering 5 follicles and a 2 mm2 reservoir on average covering 1 follicle. Covering 5 or fewer dermal channels helps improve reliability of the measurements because a low number of dermal channels provides a substantially linear (e.g. quasi-linear) relationship that can be exploited to determine the number of dermal channels N, as will be explained further herein (Figures 7a-7b). If a channel density is lower than the example of 50 channels per cm2, then larger reservoir sizes help to improve the likelihood of covering up to 5 channels. Increasing the number of reservoirs having these dimensions can also improve the likelihood of having individual reservoirs covering between 1 to 5 dermal channels. Once the number of dermal channels N is determined for a given reservoir, the flux of the target substance jtarget can be evaluated for a selected number of dermal channels n in that reservoir. The flux of the target substance can be determined based on the formula jtarqet = Ctarget reservoir v • - where V is the volume of the reservoir, / extractionis an extraction current that is applied to cause transdermal iontophoretic extraction of the target substance into the reservoir, and At is the amount of time for which the extraction current was applied. For simplicity, n may optionally be selected as 1 dermal channel. It is noted that the function in the numerator, i.e. Ctarget reservoir ■ V, is the mass of the target substance within the reservoir, and that the concentration in the reservoir can be determined based on the known values of current or voltage and sensitivity as Ctarget reServoir = or Ctarget reServoir = Utarget as discussed above. ^target Although the target concentration Ctarget subject is calculated from measurements, the reference concentration Creference-subject may instead be a pre-determined value (e.g. from a look-up table). For example, in some embodiments, the reference substance is the same as the target substance (thus, the proportionality constant K = 1). In such cases, the reference concentration Creferencesubject may correspond to an expected concentration of the target analyte in the subject at a particular (reference) flux of the target substance, as will be explained further below. In other embodiments, the reference substance is different from the target substance. For example, the reference substance may be a substance which is known to have a substantially constant concentration in the subject (e.g. a substantially constant subdermal concentration) across large groups of people, such that it can be reliably assumed to have a particular value Creferencesubject for a given person (e.g. based on race, sex, age, medical condition, etc). The sensitivity to different substances ^target and / or ^reference maY be known e.g. from manufacturer guidance. Each sensitivity may be stored in a memory of the processing unit or may be received from a remote server, e.g. using a look-up table. The reference flux jreference can be obtained at any stage in the method, e.g. before, after, or concurrently with any of the above-mentioned steps. Notably, the reference flux jreference corresponds to the flux of the reference substance through the same selected number of dermal channels n (e.g. one dermal channel) as the number of dermal channels n for which the flux of the target substance jtarget is determined. The two fluxes can therefore be easily compared, e.g. according to the relationship Ctargei subject — * Creference_subject * K. Jreference In some embodiments, obtaining the reference flux jreference comprises selecting a pre-determined substantially constant value for flux of the reference substance at the reference concentration Creference_subject- For example, the reference flux jreference may be obtained using a look-up table which provides a flux of the reference substance at a given concentration (e.g. reference subdermal concentration, e.g. 10 mM) through a given number n of dermal channels (e.g. 1 dermal channel). This allows for the method to be used in combination with a simple sensor structure, since no additional sensing is required to determine the reference flux. However, the accuracy of such embodiments relies on the reference substance having a substantially constant value for flux, across a variety of people, such that it can be reliably determined and applied to an individual subject without requiring additional measurements. For example, the reference flux may be substantially constant across a group of people having the same weight, body-mass index, race, age, sex, or medical condition (e.g. hydration level, kidney condition). Therefore, optionally, the reference flux is selected based on one or more of the following characteristics of the subject: weight, skin follicle maturation cycle, body-mass index, race, age, sex, or medical condition. For example, the look-up table may include a plurality of flux values in dependence on these characteristics. The look-up table may be stored in the memory of the processing unit. Optionally, the reference substance and the target substance may both be neutral (non-charged) substances. Thus, they may both be extracted via electro-osmosis rather than electro-migration, which helps simplify calculations since the primary contributions to flux occur through the dermal channels themselves (rather than through skin regions without dermal channels). Optionally, the reference substance is the same as the target substance. This further simplifies the estimation of concentration since the proportionality constant K will then have a value of 1. Optionally, the reference substance and target substance may both comprise glucose. Without wishing to be bound by theory, the inventors have identified evidence to suggest that glucose flux per dermal channel at a given systemic concentration may be substantially constant (e.g. within ±15%, e.g. within ±10%) in a given region (e.g. subdermal region) across particular groups of people and thus could be selected from a reference table to provide the reference flux inference- In instances where the dermal channels comprise pre-existing (naturally occurring) pathways in the subject’s skin, the flux may differ based on different characteristics of the dermal channels across people, e.g. based on differing dimensions of the dermal channels. Since the characteristics of naturally-occurring dermal channels may correlate with common characteristics (e.g. age, race, sex), a look-up table of flux values could be used to select a likely reference flux jreference based on a user’s individual characteristics. Optionally, the reference flux may be calculated based on a measurement from the sensor, rather than being selected from a look-up table. For example, optionally, the reference substance is different from the target substance; wherein obtaining the reference flux jreference through the selected number of dermal channels n comprises: receiving a reference electric signal (e.g. having a reference current Jreference or reference voltage Ureference) corresponding to an amount of the reference substance that is extracted into the reservoir of the sensor through the set of dermal channels; and determining the reference flux inference through the selected number of dermal channels n, based on the reference electric signal (e.g. reference current Inference or reference voltage ^reference), the determined number of dermal channels N, and the reference concentration of the reference substance Creference_subjeCf The reference flux may be evaluated similarly to the target flux based on the equation inference = Creference reservoir v n where reference reservoir is the concentration of the reference substance in the a j. T Ar / t. / t. / C / lLt. / CoC! l / ULl ^'‘extraction "reservoir. The reference electric signal may comprise a reference current Inference or a reference voltage ^reference that corresponds to the amount of the reference substance that is extracted into the reservoir through the set of dermal channels. For example, in embodiments utilising chronoamperometry to detect the reference substance, the reference electric signal may comprise a reference current Inference- In embodiments utilising potentiometry to detect the target substance, the reference electric signal may comprise a reference voltage Ureference- The reference substance and target substance may be detected using similar methods (e.g. both using chronoamperometry or both using potentiometry) or using different methods (e.g. with one being detected using chronoamperometry and the other being detected using potentiometry). Similarly to the target substance, the concentration of the reference substance in the reservoir may be calculated either as Creference reservoir = Or Creference reservoir = based On the knOWn "reference "reference sensitivity of the sensor to the reference substance Preference- This method assumes that the concentration of the reference substance in the relevant region of the subject’s body (e.g. subdermal region) is approximately known. Accordingly, this is particularly useful for embodiments in which the reference substance is an analyte that has a substantially constant concentration in a particular region across a large group of people and can therefore be inferred for the individual subject. This method may be particularly advantageous for reference substances that are extracted primarily via the same dermal channels as the target substance. For example, in some embodiments, the target substance is a neutral (non-charged) substance, such as glucose, , where the subdermal concentration varies in time. In some embodiments, the reference substance is also a neutral (noncharged) substance, albeit one which has a substantially constant concentration (e.g. in a subdermal region) in time and across people. Since both substances are neutral (uncharged), they will both be preferentially extracted through the dermal channels via electro-osmosis rather than also being extracted through other portions of the skin via electro-migration. This allows for a simple sensor configuration and simple processing method, since both measured fluxes inference and jtarget can correspond to the same number of dermal channels N, without requiring any additional calculations. In some embodiments, the neutral reference substance is urea. Urea is a neutral (uncharged) substance and therefore is primarily iontophoretically extracted through dermal channels via electroosmosis, rather than electromigrating through other regions of the skin. Urea has a substantially constant concentration both in time and across large groups of people (e.g. excluding those with kidney disease). A neutral substance such as urea may be detected in a simple manner using chronoamperometry. In other embodiments, it may be preferred to use a charged substance as a reference substance. A charged substance may also be detected using chronoamperometry, e.g. using an ion-selective field effect transistor (ISFET) to measure a current. Alternatively, the charged substance may be detected using potentiometry, which may provide a simpler detection arrangement. In some embodiments, the charged reference substance is Na+. Na+ also has a substantially constant concentration in time and across large groups of people (unless someone is severely dehydrated, for example). Na+ concentrations are rarely seen outside 125-140 mM (with a departure from this range typically indicating severe sickness). Due to this consistency, Na+ may be a useful reference substance to which other substances can be related. While not wishing to be bound by theory, it is believed that Na+ concentrations may be less affected by kidney disease than e.g. urea. The use of Na+ may therefore provide a further advantage, for example, in combination with glucose monitoring, since kidney disease is a comorbidity with diabetes, for which glucose monitoring is typically utilised. Accordingly, the use of Na+ can enable improved reliability of results when relying on a predetermined substantially constant concentration across groups of people. However, since Na+ is a charged molecule, it will be extracted via electromigration, and thus will be extracted over a larger portion of skin than just the dermal channels. The method may therefore be adapted to account for this variation. For example, optionally, the reference substance is different from the target substance; wherein obtaining the reference flux inference through the selected number of dermal channels n, comprises: receiving three reference electric signals (e.g. having respective reference currents ^reference_l >Ireference_2, aad Ireference_3 C respective reference VOltageS Ureference_l >^reference_2 >and ^reference_3) corresponding to amounts of the reference substance that are extracted into three reservoirs of the sensor respectively, wherein the three reservoirs each have different surface areas Slt S2, and S3; determining a total flux of the reference substance in each reservoir jreference_sr, Jreference_s2, Jreference_s3 using the reference electric signals; and comparing the total fluxes of the reference substance in each reservoir to determine the reference flux jreference through the selected number of dermal channels n. The step of comparing the total fluxes may include plotting the total fluxes Jreference_siJreference_s2, and Jreference_s3, against the reservoir areas Slt S2, and S3, and identifying a y-intercept, which may correspond to the flux inference through a single dermal channel (n = 1). This enables any flux measured from non-dermal channels (which may also be referred to as “non-channels” or areas of the skin without dermal channels) to be disregarded, thereby improving reliability of results e.g. when comparing with an uncharged (neutral) target substance, as will be explained further herein (e.g. with reference to Figures 13 and 14). As used herein, the term “surface area” may refer to a two-dimensional area of the reservoir that faces (covers, contacts) the skin in use. The differences in surface areas across the three reservoirs is useful because this allows the extraction from skin without dermal channels to be compared and accounted for, so that only the extraction from dermal channels is used when comparing with the flux of the target analyte. Surface area may be interchangeably referred to simply as an “area”. Optionally, the three reservoirs may have the same depth (as measured in a direction away from the skin in use). The processing unit may be configured to receive any of the reference signals described above from the sensor. Optionally, there is provided a method for estimating a concentration of a target substance in a subject, the method comprising: generating the primary electric signal, using the sensor, by: using a set of extraction electrodes (e.g. two or more electrodes) of the sensor to apply an extraction current to the subject’s skin to cause transdermal iontophoretic extraction of the target substance from the subject into the reservoir; and using a set of detection electrodes of the sensor to detect the primary electric signal (e.g. as a current Itarget or voltage Utarget) that corresponds to the amount of the target substance extracted into the reservoir; wherein the method further includes performing any of the computer-implemented methods described above using the primary electric signal. Optionally, the method comprises monitoring the concentration of the target substance over a set time period, by periodically (at regular or irregular intervals within the time period) repeating the steps for generating the primary electric signal (using the sensor), receiving the primary electric signal (e.g. at the processing unit), determining the flux of the target substance jtarget, and estimating the concentration of the target substance. The method can therefore be used for continuous monitoring without requiring routine invasive calibration (e.g. via a finger stick test). The sensor may be retained in the same position on the body throughout the monitoring, such that it is not required to re-evaluate the number of dermal channels N for each measurement. Optionally, there is provided a method for estimating a concentration of a target substance in a subject, the method comprising: generating reference electric signal(s), using the sensor, by: using a set of extraction electrodes of the sensor to apply an extraction current to the subject’s skin to cause transdermal iontophoretic extraction of the reference substance from the subject into one or more reservoirs of the sensor; and using a set of detection electrodes of the sensor to detect one or more reference electric signals (e.g. having the currents Inference, Inference.!, Ireference_2, and Ireference_3 or voltages Ureference, Ureference l, Ureference 2, and Ureference_3 discussed above) each corresponding to an amount of the reference substance extracted into a respective reservoir; and performing any of the computer-implemented methods discussed above using the one or more reference electric signals to determine the reference flux inference- This method may be combined with any of the methods discussed above, e.g. in combination with the method for generating the primary electric signal using the sensor. In embodiments having multiple reservoirs (e.g. three reservoirs of different areas), each reservoir may have its own set of extraction and / or detection electrodes. The extraction electrodes for applying the extraction current to the subject’s skin to extract the reference substance may be the same or different as the extraction electrodes for iontophoretically extracting the target substance. In some embodiments, one or more extraction electrodes can be configured to extract a substance into multiple reservoirs (e.g. as shown in Figure 12). Using a common extraction electrode (that serves multiple reservoirs) can help simplify the arrangement. The detection electrodes for detecting the reference substance may be configured differently (e.g. formed of different materials) from the electrodes for detecting the target substance, since each electrode may be sensitive to different analytes. In any embodiment where the reference flux is based on one or more measured reference electric signals from a sensor and the concentration is periodically monitored within a time period, optionally, the reference electrical signal(s) are generated less frequently within the time period than the primary electric signal. For example, the reference substance may only be extracted and measured at the beginning of operation. This can help avoid accumulation of the reference substance in the reservoir, which could otherwise take place after multiple instances of extraction / monitoring, particularly for substances which are extracted over a large area (e.g. charged substances such as Na+). In such embodiments, the sensor may include a first subset of sensor pixels for monitoring the target substance only, and a second subset sensor pixels for co-monitoring the reference substance and the target substance, wherein the sensor is configured to use the extraction electrodes to apply an extraction current in the second subset of sensor pixels less frequently than in the first subset of sensor pixels. According to another aspect of the invention, there is provided an apparatus for extracting and detecting a target substance from a subject, the apparatus comprising: the sensor; and a processing unit having at least one processor and at least one memory including computer program code, wherein the computer program code is configured to, with the at least one processor, cause the processing unit to perform any of the computer-implemented methods discussed above. The sensor may have one or more sensor “pixels”, each sensor pixel having: a substrate portion for positioning on the subject’s skin, a reservoir formed in the substrate portion, and a set of electrodes operatively coupled to the reservoir. In each sensor pixel, the set of electrodes may include one or more target detection electrodes configured to detect a current Itarget or voltage Utarget that corresponds to an amount of the target substance extracted into the respective reservoir. Optionally, the set of electrodes includes one or more extraction electrodes configured to apply an extraction current to the subject’s skin to cause transdermal iontophoretic extraction of one or more substances including the target substance (and optionally the reference substance) from the interstitial fluid into the reservoir. In embodiments having only one extraction electrode in a pixel, an adjacent pixel may have a further extraction electrode, wherein the two extraction electrodes from the respective pixels are mutually configured to together apply the extraction current to their respective reservoirs. In other embodiments, a given pixel may have a set of (e.g. two or more) extraction electrodes within the same pixel. Optionally, the set of electrodes further includes one or more reference detection electrodes configured to detect a current Ireference or voltage Ureference that corresponds to an amount of the reference substance extracted into the respective reservoir. Optionally, each set of target and / or reference detection electrodes may be configured for chronoamperometric detection, e.g. by having a respective Working electrode, Counter electrode, and Reference electrode. In some embodiments, the Reference and Counter electrodes may be combined in one. Optionally, each set of target and / or reference detection electrodes may be configured for potentiometry, e.g. by having an ion-selective electrode and a Reference electrode. Optionally, the sensor may comprise one or more channels (“artificial” dermal channels e.g. microneedles) for extending through the skin and extracting the one or more substances into the reservoir. In embodiments in which the sensor comprises the dermal channels (artificial dermal channels such as microneedles), the sensor may be configured to perform iontophoretic extraction to extract the substance(s) through the dermal channels, in a similar manner as discussed elsewhere herein. Alternatively, the sensor may not be configured to perform iontophoretic extraction. Instead, the sensor may simply draw the substance(s) through the artificial dermal channels in any other known method, e.g. simply relying on capillary action. Optionally, the processing unit may include a controller configured to control the sensor to detect (e.g. periodically detect) the reference and / or primary electric signals. The sensor may then relay the primary / reference electric signals to the processing unit for processing. Optionally, the sensor may comprise a plurality of said sensor pixels. In such embodiments, the processing unit may be configured to compare a plurality of electric signals extracted via the detection electrodes to quantise a value (e.g. current, mass, or voltage value) corresponding to an amount of substance per dermal channel, as discussed above. A plurality of sensor pixels may therefore be particularly advantageous for the non-invasive arrangements which utilise pre-existing dermal channels in the subject’s skin (e.g. skin appendages), rather than arrangements in which the sensor itself comprises an already-known number of dermal channels (e.g. microneedles). The processing unit may be physically attached to the sensor and configured to communicate therewith (e.g. via a wired communication channel). Alternatively, the processing unit may be physically detached and remote from the sensor as a separate device (e.g. mobile device) and may communicate with the sensor via a wireless communication channel. According to an aspect of the invention, there is provided: a sensor for extracting and detecting a target substance from a subject, the sensor comprising: a sensor pixel assembly comprising a comonitoring sensor pixel having: a substrate portion for positioning on the subject’s skin, the substrate portion comprising a reservoir assembly having one or more reservoirs; and a set of electrodes including: one or more extraction electrodes configured to apply a current to the subject’s skin to cause transdermal iontophoretic extraction of the target substance and a reference substance into each reservoir of the one or more reservoirs; one or more target detection electrodes configured to detect a primary electric signal (e.g. having a target current or target voltage) that corresponds to an amount of the target substance extracted into each reservoir of the one or more reservoirs; and one or more reference detection electrodes configured to detect a reference electric signal (e.g. having a reference current or reference voltage) that corresponds to an amount of the reference substance extracted into each reservoir of the one or more reservoirs. The sensor according to this aspect may include any of the features already discussed above. This co-monitoring sensor pixel can be particularly advantageous for monitoring a target substance using a measured reference flux. For example, optionally, the sensor may be provided in combination with a processing unit that is configured to perform any method discussed above in which the reference substance is different from the target substance and in which the reference flux jreference is evaluated based on one or more measured signals (e.g. currents, voltages) relating to the reference substance (rather than e.g. based on a look-up table). The co-monitoring sensor pixel may also be referred to as a “calibration sensor pixel”. The comonitoring sensor pixel, extraction electrodes, reference detection electrodes, and target detection electrodes, may have any of the features already discussed above. The substrate may comprise a flexible layer (e.g. elastomer). Each reservoir may comprise a gel (e.g. agarose). The gel may have an enzyme for reacting with the reference and / or target substances to facilitate detection thereof. The enzyme may provide the sensor pixels with specificity of responses to the target and / or reference substances. For example, for detecting glucose, the enzyme glucose oxidase may be entrapped in the gel. In other embodiments, the enzyme may be entrapped on the working electrode rather than in the gel. In other embodiments, the substance(s) may be detected in a non-enzymatic manner, without requiring any enzyme. Further examples of possible configurations for the sensor structure (including e.g. possible materials for the reservoirs and electrodes) may be found in WO 2017 / 186783, which is incorporated herein by reference in its entirety. The reservoirs may be physically separate from each other. The reservoirs may be formed within the substrate, e.g. partially or completely embedded within the substrate. Optionally, the reservoir assembly includes three reservoirs having different surface areas. Having reservoirs of different areas enables the measurements obtained via dermal channels to be determined and separated from the measurements obtained via non-dermal channels. This can be particularly advantageous for monitoring target and reference substances that are extracted via different pathways in the skin, e.g. with one substance being extracted primarily via dermal channels (e.g. a neutral substance such as glucose which is extracted via electro-osmosis) while another substance is extracted across all regions of the skin (e.g. a charged substance, e.g. an ion such as Na+ which is extracted via electro-migration). In such an arrangement, the set of electrodes may include a respective set of detection electrodes for each reservoir to detect the signals that correspond to an amount of the target substance and reference substance in each reservoir. In some embodiments, the reservoir assembly may also include further reservoirs, providing three or more reservoirs which may each have different areas. However, providing only three reservoirs of different areas may help to balance a simple and compact construction with providing the desired measurements. Optionally, the areas of the three reservoirs are non-divisible by each other. Optionally, the areas of the three reservoirs are in a ratio of prime numbers. This helps improve accuracy during analysis, with prime number ratios providing the greatest improvement in accuracy. Optionally, the sensor includes a plurality of extraction electrodes including an inner extraction electrode and one or more outer extraction electrodes separated by a gap. The inner extraction electrode may operate as an anode. The one or more outer extraction electrodes may each operate as a cathode. Optionally, the outer extraction electrode(s) are arranged around a periphery of the inner extraction electrode. Optionally, the outer extraction electrode(s) are arranged between the inner extraction electrode and the target detection electrode(s). Optionally, the outer extraction electrode(s) are arranged between the inner extraction electrode and the reference detection electrode(s). Optionally, each outer extraction electrode may be configured to cause iontophoretic extraction into a single reservoir, which may have only a single outer extraction electrode associated therewith. This can provide a simplified and compact construction. Optionally, the outer extraction electrode(s) may partition the reference detection electrode(s) from the target detection electrode(s). As used herein, the terms “inner” and “outer” may refer to different regions of the sensor across its surface that faces the skin in use, e.g. with the “inner” extraction electrode being closer to a centre of the sensor pixel in use (away from the lateral periphery of the sensor pixel) and the “outer” extraction electrode being nearer to a periphery of the sensor pixel in use, where both the “inner” extraction electrode and “outer” extraction electrode may be located at substantially the same depth within the sensor (same distance from skin in use). Optionally, the inner extraction electrode (or a majority thereof) may be symmetrically arranged with respect to the outer extraction electrodes. Optionally, the inner extraction electrode (or a majority thereof) may be substantially equidistant from each of the outer extraction electrodes. This can help establish the same electric field between the inner extraction electrode and each of the outer extraction electrodes. Optionally, the reservoirs are each separated by a gap. As used herein, a “gap” may refer to a region of the sensor (e.g. sensor substrate) into which the target / reference substances are not extracted and / or not detected. Optionally, the target detection electrode(s) may comprise a working electrode and a counter and reference electrode. Optionally, the reference detection electrode(s) may comprise a working electrode and a counter and reference electrode. The term “reference electrode” in this context takes its normal meaning as referring to an electrode that has a stable and well-known electrode potential. The skilled person will understand this is not limited to an electrode for detecting the “reference substance” (which may alternatively be referred to as a “secondary substance” or “comparison substance” or “standardisation substance”). Optionally, the co-monitoring pixel is round (e.g. circular). Optionally, one or more of the one or more extraction electrodes (e.g. the inner extraction electrode) is round (e.g. circular). Optionally, the three reservoirs are arranged around a periphery of at least one of the one or more extraction electrodes (e.g. around an inner extraction electrode). This provides a compact shape and may help establish substantially the same electric field between the extraction electrode and each of the reservoirs. Optionally, the three reservoirs may together form a ring around the extraction electrode(s) (e.g. around the inner extraction electrode). Optionally, the sensor pixel assembly comprises one or more additional sensor pixels, each additional sensor pixel having: an additional substrate for positioning on the subject’s skin, the additional substrate comprising an additional reservoir assembly having one or more additional reservoirs; and a set of additional electrodes including: an additional extraction electrode configured to apply a current to the subject’s skin to cause transdermal iontophoretic extraction of the target substance from the interstitial fluid into each additional reservoir of the one or more additional reservoirs; and one or more additional target detection electrodes configured to detect a primary electric signal (e.g. having a target current or target voltage) that corresponds to an amount of the target substance extracted into each additional reservoir of the one or more additional reservoirs. Having a plurality of sensor pixels allows the number of dermal channels at a given reservoir to be statistically determined, e.g. using the method described above. The additional substrate portion of each additional sensor pixel may be integrally formed with the substrate portion of the co-monitoring sensor pixel. Thus, the sensor may comprise a substrate (single substrate) which includes a plurality of continuous substrate portions for the plurality of sensor pixels. Optionally, the additional sensor pixel(s) may comprise target sensor pixel(s) (target-only sensor pixel(s)). As used herein, the term “target sensor pixel” or “target-only sensor pixel” may refer to a sensor pixel that is not configured to detect a reference electric signal that corresponds to an amount of reference substance extracted into its respective reservoir. Accordingly, such a pixel may omit the reference detection electrodes. Using target sensor pixels may therefore help to provide a compact configuration. Optionally, the additional sensor pixel(s) may be additional co-monitoring sensor pixel(s). In such embodiments, the additional extraction electrode(s) may be configured to also cause transdermal iontophoretic extraction of the reference substance from the interstitial fluid into each additional reservoir of the one or more additional reservoirs, and the additional co-monitoring sensor pixel may further comprise one or more additional reference detection electrodes configured to detect a reference electric signal (e.g. having a reference current or reference voltage) that corresponds to an amount of the reference substance extracted into each additional reservoir of the one or more additional reservoirs of the respective additional co-monitoring sensor pixel. The additional co-monitoring sensor pixel(s) may have any of the features discussed above in relation to the co-monitoring sensor pixel. By providing more than one co-monitoring sensor pixel, the accuracy of the measurements can be improved, since the reference substance can be extracted across a larger number of pixels, thereby providing a larger amount of data that can be used in determining the concentration of the target substance (compared to other embodiments which may assume that the reference substance measurements would be the same for all pixels, and thus may utilise this assumption in calculating the concentration of the target substance in the subject). Optionally, a majority (e.g. all) of the sensor pixels are configured as co-monitoring sensor pixels. Each co-monitoring sensor pixel may be configured in the manner discussed above, i.e. having an extraction electrode that is further configured to cause transdermal iontophoretic extraction of the reference substance, and further having one or more reference detection electrodes configured to detect a reference electric signal that corresponds to an amount of the reference substance extracted into each reservoir of the co-monitoring pixel. In arrangements where the target substance and reference substance are both extracted primarily through dermal channels (e.g. glucose and urea), the majority (e.g. all) of the sensor pixels can be configured as co-monitoring sensor pixels without significantly increasing the complexity of the device, since the co-monitoring pixels in this case may only include a single reservoir. Alternatively, optionally, only a minority (e.g. one) of the sensor pixels may be configured as a comonitoring sensor pixel(s). For example, the majority of the sensor pixels may be configured as target sensor pixels (target-only sensor pixels), which may omit reference detection electrodes. This may be particularly useful to provide a simplified and compact sensor arrangement, e.g. in embodiments where the target substance and reference substance are iontophoretically extracted in different manners (e.g. glucose and Na+) and which therefore may utilise a more complex co-monitoring arrangement (e.g. having three reservoirs as discussed above). In such embodiments, the comonitoring sensor pixels may be relatively large (due to the plurality of reservoirs) and so it may be advantageous to limit the number of co-monitoring sensor pixels present, so as to provide an overall compact form factor. The co-monitoring sensor pixels may provide information about representative dermal pathways on the particular patch of skin interrogated by the co-monitoring sensor pixel, which can then be extrapolated to other target-only pixels in the array. Optionally, the (or each) target detection electrode is configured to detect a neutral (uncharged) target substance present in its respective reservoir. This may be achieved by forming the detection electrode(s) from particular materials, using known methods. Additionally, the reservoir may contain an enzyme (e.g. in a gel) to react with the target substance to produce a signal (e.g. current) that is detected by the target detection electrodes in use. Optionally, the (or each) reference detection electrode is configured to detect a charged (non-neutral) reference substance present in its respective reservoir. Alternatively, optionally, the (or each) reference detection electrode is configured to detect a neutral (uncharged) reference substance. As above, each reference detection electrode may be configured to detect a specific reference substance by forming the reference detection electrode(s) from particular materials, using known methods. The reservoir may also contain an enzyme (e.g. in a gel) to react with the reference substance to produce a signal (e.g. current or voltage) that is detected by the reference detection electrodes in use. Other methods of detection include providing an enzyme immobilised on the (target or reference) detection electrode, or performing non-enzymatic detection. Optionally, any of the sensors described above can be provided in combination with a processing unit configured to perform any of the computer-implemented methods discussed above. The invention includes the combination of the aspects and preferred features described except where such a combination is clearly impermissible or expressly avoided. Summary of the Figures Embodiments and experiments illustrating the principles of the invention will now be discussed with reference to the accompanying figures in which: Figure 1 shows a a cabinet projection of an embodiment sensor for extracting and detecting a target substance from a subdermal region of a subject’s skin. Figure 2 shows a cabinet projection of the sensor of Figure 1, applied to the subject’s skin to overlie naturally-occurring dermal channels in the subject’s skin. Figures 3a and 3b shows schematic drawings of processing units according to embodiments of the invention. Figure 4 shows a flow chart of a computer-implemented method for estimating a concentration of a target substance according to an embodiment of the invention. Figure 5 shows a flow chart of a computer-implemented method for determining a number of dermal channels in a reservoir according to an embodiment of the invention. Figures 6a to 6c show graphs of representative data useful to help illustrate the method steps of Figure 5. Figures 7a to 7b show modelling results indicating a substantially linear (e.g. quasi-linear) relationship between current and number of dermal channels at a small number of dermal channels (e.g. up to five dermal channels), useful for understanding the method of Figure 5. Figure 8 shows experimental data of measured current over time as extracted and detected through four sensor pixels each covering different numbers of dermal channels. Figures 9 shows a binning operation performed on experimental data similar to those provided by Figure 8 useful for identifying a linear relationship and thus quantising an amount of target substance extracted through a single dermal channel, and then measured, in a gel reservoir. Figure 10 shows a cabinet projection of a sensor according to another embodiment of the invention, which includes at least one co-monitoring pixel. Figure 11 shows a top-down view of an embodiment co-monitoring pixel which may be particularly useful for detecting substances that are iontophoretically extracted primarily through electro-osmosis. Figure 12 shows a top-down view of an embodiment co-monitoring pixel having three sub-regions of different areas, which may be particularly useful for detecting a substance (e.g. a reference substance) that is iontophoretically extracted primarily through electro-migration. Figure 13 shows a schematic diagram useful for understanding the theory behind the sub-regions of different areas in Figure 12. Figures 14 shows a graph of measured flux against area, useful for demonstrating a method that can be performed to evaluate a reference flux through a single a dermal channel using the co-monitoring pixel of Figure 12 and with the assumptions from Figure 13. Detailed Description of the Invention Aspects and embodiments of the present invention will now be discussed with reference to the accompanying figures. Further aspects and embodiments will be apparent to those skilled in the art. All documents mentioned in this text are incorporated herein by reference. Sensor Figure 1 shows an embodiment sensor 100 which is configured to extract and detect a target substance from a subject (e.g. from the interstitial fluid of a subject). The sensor 100 has a substrate 102 which, in this embodiment, includes a first layer 104 and a second layer 106. The first layer may be formed of an elastomer, such as polydimethylsiloxane (PDMS). The second layer may be formed of a flexible material such as polyethylene terephthalate (PET), polyethylene naphtalate (PEN), or a polyimide film such as Kapton. Further examples of possible configurations for the substrate 102 may be found in WO 2017 / 186783, which is incorporated herein by reference in its entirety. The substrate 102 comprises a plurality of substrate portions, each substrate portion having a respective reservoir 108a-d and belonging to a respective sensor pixel 110a-d. In Figure 1, the sensor 100 includes four sensor pixels 110a-d arranged in a 2x2 array. In variant embodiments, the sensor 100 may comprise a different number of sensor pixels 110a-d, or may have the sensor pixels arranged in a different pattern and / or a different geometrical relationship relative to each other. Each sensor pixel 110a-d further includes a set of electrodes, which are formed in the second layer 106 of the substrate 102. For simplicity the electrodes are labelled in Figure 1 only with respect to the set of electrodes 112a of the first sensor pixel 110a. The remaining pixels 100b-d also have corresponding electrodes. The set of electrodes 112a includes one or more extraction electrodes for extracting a target substance into the reservoir 108a, and one or more detection electrodes for detecting a target electric signal that corresponds to the amount of the target substance extracted into the reservoir 108a. The one or more detection electrodes may include a working electrode and reference electrode, e.g. an Ag working electrode and an Ag / AgCI reference electrode for a target substance of glucose. Alternatively, the one or more detection electrodes may include a working electrode (e.g. graphene with Pt nanoparticles, when glucose is the target substance), a reference electrode (which may be the same reference electrode as used for extraction), and a counter electrode (e.g. Pt counter electrode). Other configurations are possible, for example as described in WO 2017 / 186783. Each reservoir 108a-d is substantially cylindrical. In variant embodiments, the reservoirs may have different shapes. Each reservoir 108a-d contains a polymer matrix into which target substances are collected from the skin and dermal channels. A biorecognition element (e.g. enzyme) may be incorporated into the polymer matrix of the reservoir to react with a target substance to produce the analyte which is then detected by the detection electrodes. Alternatively, the biorecognition element may be included in a hybrid material intimately connected to the detection electrode(s) (e.g. a working electrode). In other examples, each reservoir may contain a gel having an enzyme therein for reacting with the target substance. For example, the enzyme glucose oxidase may be incorporated in the gel to react with a target substance of glucose to produce hydrogen peroxide which is then detected by the detection electrodes. Figure 2 shows the sensor 100 positioned in use over a subject’s skin 200. Beneath the surface of the skin is a subdermal region 202 having interstitial fluid, in which the target substance (e.g. glucose) is naturally found. In this embodiment, the sensor 100 does not include dermal channels therein, e.g. it does not include needles for inserting into the skin 200 to extract the target substance from the subdermal region 202. Instead, the sensor 100 is an entirely non-invasive device which is only located externally to the skin 200 in use. The sensor 100 therefore transdermally extracts the target substance from naturally-occurring dermal channels 204, which are shown in Figure 2 as skin appendages in the form of hair follicles. In use, with the sensor applied over the skin 200, each reservoir 108a-d may be aligned over a different number of dermal channels 204. Each reservoir 108a-d may have a surface area (at a flat end face configured to contact the skin in use) of at least 2 mm2 and up to 15 mm2, optionally up to 10 mm2. When used on human skin 200, these dimensions have been shown to provide good reliability in overlying between 1 and 5 hair follicles on average per reservoir. In this regard, it is noted that Figure 2 is not drawn to scale. In use, each sensor pixel is configured to use its extraction electrode(s) to apply an extraction current to the subject’s skin 200 to cause transdermal iontophoretic extraction of the target substance from the subdermal region 202 through any dermal channels 204 aligned with its respective reservoir 108a-d. In some examples, once the target substance is drawn into the reservoir 108a-d, it may optionally react with an enzyme in the reservoir to produce a chemical or electron that is then detected by the detection electrode(s) in the form of a signal having a current. In other embodiments, the target substance may be detected in other manners, e.g. non-enzymatically. In this embodiment, the sensor 100 does not include any co-monitoring pixels for detecting a reference electric signal that corresponds to an amount of a reference substance received in the reservoir. Instead, the sensor pixels 110 are all configured as target-only sensor pixels, which only measure the current corresponding to an amount of target substance (e.g. glucose) in the reservoir. This information is then sent to a processing unit which uses the measured electric signal to estimate a subdermal concentration of the target substance, by comparison with a reference flux value which would be expected for a reference substance that is the same as the target substance (e.g. glucose) at a reference subdermal concentration. Processing unit Figure 3a shows a schematic drawing of a processing unit 300 according to embodiments of the invention. The processing unit has at least one processor 302 and at least one memory 304, and a sensor interface 306. The sensor interface 306 is configured to receive signals from a sensor (e.g. the sensor 100), e.g. in a wired or wireless manner. The at least one memory includes computer program code. The computer program code is configured to, with the at least one processor 302, cause the processing unit to perform a computer-implemented method to estimate the concentration of the target substance in the subject. Figure 3b shows a schematic drawing of another processing unit 3000 according to embodiments of the invention. The example processing unit 3000 includes a processor 3004 for executing software routines. Although a single processor is shown for the sake of clarity, the processing unit 3000 may also include a multi-processor system. The processor 3004 is connected to a communication infrastructure 3006 for communication with other components of the processing unit 3000. The communication infrastructure 3006 may include, for example, a communications bus, cross-bar, or network. The processing unit 3000 further includes a main memory 3008, such as a random-access memory (RAM), and a secondary memory 3010. The secondary memory 3010 may include, for example, a hard disk drive 3012 and / or a removable storage drive 3014, which may include an optical disk drive, solid state storage or the like. The removable storage drive 3014 reads from and / or writes to a removable storage unit 3018 in a well-known manner. The removable storage unit 3018 may include an optical disk, removable solid-state storage (e.g. SD card) or the like, which is read by and written to by removable storage drive 3014. As will be appreciated by persons skilled in the relevant art(s), the removable storage unit 3018 includes a computer readable storage medium having stored therein computer executable program code instructions and / or data. In an alternative implementation, the secondary memory 3010 may additionally or alternatively include other similar means for allowing computer programs or other instructions to be loaded into the processing unit 3000. Such means can include, for example, a removable storage unit 3022 and an interface 3020. Examples of a removable storage unit 3022 and interface 3020 include a program cartridge and cartridge interface (such as that found in video game console devices), a removable memory chip (such as an EPROM or PROM) and associated socket, and other removable storage units 3022 and interfaces 3020 which allow software and data to be transferred from the removable storage unit 3022 to the processing unit 3000. The processing unit 3000 also includes at least one communication interface 3024. The communication interface 3024 allows software and data to be transferred between processing unit 3000 and external devices (e.g. the sensor) via a communication path 3026. In various embodiments, the communication interface 3024 permits data to be transferred between the processing unit 3000 and a data communication network, such as a public data or private data communication network. The communication interface 3024 may be used to exchange data between a plurality of different processing unit 3000 that together form an interconnected computer network. Examples of a communication interface 3024 can include a modem, a network interface (such as an Ethernet card), a communication port, an antenna with associated circuitry and the like. The communication interface 3024 may be wired or may be wireless. Software and data transferred via the communication interface 3024 are in the form of signals which can be electronic, electromagnetic, optical, or other signals capable of being received by communication interface 3024. These signals are provided to the communication interface via the communication path 3026. The processing unit 3000 further includes a display interface 3002 which performs operations for rendering images to an associated display 3030 and an audio interface 3032 for performing operations for playing audio content via associated speaker(s) 3034. As used herein, the term “computer program product” may refer, in part, to removable storage unit 3018, removable storage unit 3022, a hard disk installed in hard disk drive 3012, or a carrier wave carrying software over communication path 3026 (wireless link or cable) to communication interface 3024. These computer program products are devices for providing software to the processing unit 3000. A computer readable medium can include magnetic media, optical media, or other recordable media, or media that transmits a carrier wave or other signal. The computer programs (also called computer program code) are stored in main memory 3008 and / or secondary memory 3010. Computer programs can also be received via the communication interface 3024. Such computer programs, when executed, enable the processing unit 3000 to perform one or more features of embodiments discussed herein. In various embodiments, the computer programs, when executed, enable the processor 3004 to perform features of the embodiments discussed herein. Accordingly, such computer programs represent controllers of the processing unit 3000. Software may be stored in a computer program product and loaded into the processing unit 3000 using the removable storage drive 3014, the hard disk drive 3012, or the interface 3020. Alternatively, the computer program product may be downloaded to the processing unit 3000 over the communications path 3026. The software, when executed by the processor 3004, causes the processing unit 3000 to perform functions of embodiments described herein. It is to be understood that the embodiment of Figure 3b is presented merely by way of example. Therefore, in some embodiments one or more features of the processing unit 3000 may be omitted. Also, in some embodiments, one or more features of the processing unit 3000 may be combined. Additionally, in some embodiments, one or more features of the processing unit 3000 may be split into one or more component parts. It will be appreciated that the elements illustrated in Figure 3b function to provide means for performing the various functions and operations of the computer-implemented methods described herein. Method Figure 4 shows a flow chart of an embodiment computer-implemented method 400 for determining a concentration of a target substance in a subject. The method may be implemented by the processing unit 300. In step 402, the processing unit 300 receives a primary electric signal corresponding to (e.g. having a current Itarget or voltage Vtarget corresponding to) an amount of target substance extracted into a reservoir of a sensor through a set of dermal channels. The current Itarget or voltage Utarget may initially be measured by a sensor (e.g. sensor 100) before it is sent to the processing unit 300. In step 404, the processing unit 300 determines a number of dermal channels N in the set of dermal channels. In some embodiments, the number of dermal channels N may be pre-determined, e.g. if the sensor itself has dermal channels integrated therein (e.g. as one or more microneedles). In such embodiments, the number of dermal channels N may be obtained simply, e.g. from the memory 304 of the processing unit 300, or by receipt from the sensor (e.g. sensor 100). Thus, in such embodiments, the number of dermal channels N may be determined prior to or concurrently with receiving the primary electric signal of step 404. In other embodiments, the number of dermal channels N may require calculating, e.g. if the target substance is extracted through naturally-occurring dermal channels such as hair follicles. Even in embodiments that utilise artificially-created dermal channels (e.g. microneedles), optionally it may still be useful to calculate the number of dermal channels N rather than relying on a pre-determined figure, in case any of the microneedles are not used for extraction (e.g. due to not being inserted in the skin). An embodiment method for calculating the number of dermal channels N used for extraction will be explained further below in relation to Figures 5 to 7b. Once the processing unit 300 knows both the number of dermal channels N, and the primary electric signal corresponding to an amount of target substance extracted through said dermal channels, the processing unit may proceed to step 406 to determine a flux of the target substance jtarget through a selected number n of dermal channels. This may be determined using the equation: _ ^target reservoir ' where the concentration of the target substance in the reservoir can be determined as Ctarget reservoir = or Ctarget reservoir = (with the sensitivity otarget being in relevant units of ^target "^target current or voltage respectively). The reservoir volume V, extraction time At, and / or extraction current lcxtractionma\ be pre-stored in the memory 304 or may be received from the sensor. Returning to Figure 4, the processing unit 300 further obtains a reference flux inference of a reference substance through the selected number of dermal channels n at a reference concentration Creferenee-subjeet of the reference substance. In some embodiments, the reference substance may be the same as the target substance. In such embodiments, the reference flux jreference may be selected as an expected flux for the target substance at a given subdermal concentration, through a selected number of dermal channels. For example, the inventors have identified the following data for glucose (shown in Table below), which provides an average glucose flux across a population of users, of 2.7 nmol h1 mA1 per hair follicle (i.e. n = 1) at 10mM subdermal glucose (i.e. Creference_subject = lOmM). Set 1 set 2 set 3 set 4 set 5 set 6 set 7 set 8 Single follicle Flux (nmol h-1 mA-1) at 10mM subdermal glucose 2.9 2.9 2.3 2.9 2.9 2.5 2.5 2.4 No. of follicles 2 2 2 2 1 2 4 4 Mean = 2.7 Standard deviation = 0.26 By performing additional routine experimentation, the reference flux value may be further refined. For example, different users may have different characteristics fortheir naturally-occurring dermal channels such as hair follicles (e.g. different hair and follicle dimensions), which may affect flux. Therefore selecting the reference flux according to a look-up table based on different user characteristics may further improve accuracy. It will of course be understood that reference fluxes could similarly be obtained for other target / reference substances and could likewise be provided e.g. in a look-up table based on user characteristics. In other embodiments, reference flux may be evaluated differently, e.g. based on a reference electric signal (e.g. measured reference current or reference voltage) corresponding to an amount of a reference substance extracted into the reservoir, with the reference substance being different from the target substance. This will be explained further below in relation to Figures 10 to 14. In any case, once the reference flux is obtained, the processing unit may proceed to step 410 to estimate a subdermal concentration of the target substance based on the flux of the target substance jtarget, reference flux jreference , and reference subdermal concentration Creference_subject- This may be estimated using the equation: _ Jtarget L‘target_subject • L‘reference_subject n Jreference In embodiments where the reference substance is the same as the target substance, K = 1. In embodiments where the reference substance is different from the target substance, K may be a value other than 1. The value of K may be stored e.g. in the memory 304 of the processing unit 300 or may be received at the processing unit 300 from a remote server. Figure 5 shows a flow chart of one method for determining a number of dermal channels N according to step 404 of Figure 4. Figure 5 shows a method for embodiments in which the number of dermal channels N is calculated by the processing unit 300 rather than being pre-determined, due to the use of naturally-occurring dermal channels in the subject’s skin. In step 412, the processing unit 300 obtains a plurality of electric signals each corresponding to (e.g. having a current that corresponds to) an amount of a substance that is extracted into a reservoir of a plurality of reservoirs of the sensor. In some embodiments, this may correspond to an amount of the target substance. For example, the sensor 100 may use the extraction electrode(s) in the plurality of sensor pixels 110a-d to apply a respective extraction current to the subject’s skin 202 to cause transdermal iontophoretic extraction of the target substance from the subject into the reservoir 108a-d of the respective pixel 110a-d. Using the detection electrodes, the sensor 100 may detect a current Itarget for each respective pixel 110a-d and generate a plurality of (primary / target) electric signals to be sent to the processing unit. In other embodiments, e.g. embodiments with a plurality of co-monitoring pixels, the received plurality of signals may comprise reference electric signals (which may each have a reference current Ireference or reference voltage Ureference) which each correspond to an amount of another (reference) substance that is extracted into a respective reservoir of the plurality of reservoirs of the sensor. In such embodiments, the reference electric signals may optionally be used to determine the number of dermal channels. Once the processing unit receives these signals in step 412, it can proceed to step 414 to compare the plurality of electric signals to quantise a value (e.g. current value) corresponding to (e.g. indicative of) an amount of the substance per dermal channel (e.g. a current per dermal channel). To perform this quantisation, the processing unit can perform a method including: separating the received data into bins, performing a fitting procedure on the binned data, and solving for a current per dermal channel Io e.g. using a standard nonlinear least squares solver. Figures 6a to 9 further help illustrate this method. Figures 6a to 9 are discussed in relation to electric signals which provide a current Itarget- However, it will be recognised that, in other embodiments, the received electric signals may instead provide a voltage Utarget, in which case an equivalent method may still be used (by replacing the current values discussed in relation to Figures 6a to 6c with voltage values). Figures 6a to 6c show example representative data to help illustrate this in more detail, according to one embodiment of the invention. Figure 6a shows a set of received data points, which indicate received values for Itarget from a plurality (e.g. eight) pixels in a sensor. At this stage, the number of dermal channels (e.g. hair follicles) used for extraction into each sensor pixel is unknown. Nonetheless, as shown in Figure 6a, the received currents cluster around certain integer values. In this example, the illustrative values are: 0.1,0.9, 1.05, 1.2, 1.85, 2, 3.85, and 4.1. It can be assumed that this signal as a function of the number of dermal channels can be represented by a sequence of sharp peaks around the number of dermal channels, i.e. according to the following equation: y = Co + Io N ■ An ■ Fn n,w) n=l 0 (1) Here Co is a signal offset (corresponding to zero dermal channels in a pixel), / 0 is a signal per dermal channel (n = 1), N is the number of follicles per pixel, AN is either 0 or 1 depending on whether there is a pixel with N dermal channels, and Fw(x,w) is a peak function which is centred around N and has width w. This function could simply be a Gaussian (it should have amplitude 1 at its peak). Figure 6b shows a representation of this function, generated using equation (1), and corresponding to the illustrative situation given as an example in Figure 6a. In this example A1 = A2 = A4 = 1 and A3 = 0 (there are pixels with 1,2 and 4 dermal channels respectively, while there are no pixels with 3 dermal channels). The upper bound on the number of dermal channels can be informed by test data on the particular sensor. The sensor pixels (e.g. reservoirs) may be dimensioned to provide a good probability that most (or all) sensor pixels cover less than or equal to five dermal channels. Based on an initial estimate of / 0, denoted as 4, the processing unit can roughly classify the data points by an expected number of pixels according to the quantity Itarfet. is an initial, trial value to start a fitting procedure based on equation (2) below. This can be a generic value obtained from previous experiments involving various cohorts of people. Classifying the data points in this manner leads to a substantially linear (e.g. quasi-linear) relationship e.g. as shown in Figure 6c. The linear relationship holds well for pixels using less than or equal to five dermal channels. For pixels overlying 6 or more dermal channels, the linear relationship begins to break down, as shown in separate data of Figures 7a-7b obtained from modelling. Returning to Figures 6a-6c, in this example it can be assumed, based on the binning of the data, that the received signals correspond to sensor pixels overlying 0,1,2 and 4 dermal channels. A fitting procedure may be performed to produce a line of best fit. This may be achieved using the following function, which is of the form shown in equation (1) y= Co +id F - 1,w) + 2-F w) + 3-F (^--3, w)\ (2) Where y = x = the measured signal values from Fig 6a, i.e: X y 0.1 0.1 0.9 0.9 1.05 1.05 1.2 1.2 1.85 1.85 2 2 3.85 3.85 4.1 4.1 Equation (2) has now three unknowns, Co, / 0 and w and therefore needs at least three independent measurements to be solvable. These parameters can be solved for a standard Nonlinear Least Squares solver such as Newton-Raphson (or something more advanced). Initial values for the parameters (to help convergence) may be heuristics based on test data. If the fit using equation (2) corresponding to the proposed binning from Figure 2 is weak, the experimental values can be binned differently, generating a new equation (2) corresponding to new binning, and the fitting procedure restarted. The procedure may be stopped when the fit gives w (the width of the Gaussian functions) values that indicate clustering around integer numbers of follicles. The example given above and shown in Figures 6a-6c contains illustrative data from 8 pixels, but it can be performed for data from any number of pixels. Figures 7a and 7b show graphs of data obtained from modelling showing the relationship between measured current and number of follicles. As can be seen from these figures, at low numbers of dermal channels (e.g. less than or equal to 5 follicles, e.g. less than or equal to 4 follicles), there is a strong linear relationship between the number of dermal channels and the current carried by them. At larger numbers of dermal channels, the linear relationship begins to be lost. For this reason, it may be advantageous to dimension the sensor pixels so that the plurality of reservoirs is unlikely to each cover more than 5 follicles. Provision of a large number of sensor pixels can also help increase the likelihood of data points that relate to between 1 and 5 dermal channels, and result in a strong linear fit based on equations (1) and (2). Optionally, the quantisation method may include (e.g. as part of the binning step of Figure 6b), a step of disregarding data for bins relating to 6 or more dermal channels. Figures 8, and 9 show experimental data demonstrating the quantisation relationship. Figure 8 shows a graph of measured current overtime for a number of sensor pixels. The graph includes four curves labelled A-D, where curve A shows the measured current from a sensor pixel overlying four dermal channels (e.g. four hair follicles), curve B shows the measured current from a sensor pixel overlying three dermal channels (e.g. three hair follicles), curve C shows the measured current from a sensor pixel overlying 2 dermal channels (e.g. two hair follicles), and curve D shows the measured current from a sensor pixel overlying one dermal channel (e.g. one hair follicle). In this embodiment, the current corresponds to an amount of glucose extracted into the reservoir of each respective pixel. Figure 9 shows several sets of data similar to those shown in Figure 8, binned into different categories, by integrating and normalising each of the respective curves from graphs similar to that of Figure 8; this yields values of extracted / detected glucose. As can be seen from Figure 9, there is a clear increase in the amount of extracted glucose (and corresponding concentrations in the respective gel reservoirs) as the number of hair follicles increases. This can be evaluated (e.g. using a line of best fit) to quantise the current for a single dermal channel, Io, and the corresponding concentration in the gel reservoir. Returning to Figure 5, once the amount of current for a single dermal channel Io is determined, the method can proceed to step 416 to determine the number of dermal channels N for any given reservoir of a sensor pixel based on the quantised value. For example, N can be determined using a ratio of the current obtained from said reservoir and the quantised current per dermal channel, i.e. by taking a ratio of Alternatively, since the current is proportional to mass of the substance in the reservoir (which can be determined e.g. based on the relationship m = CV), the number of dermal channels in a given pixel can equivalently be determined as: _ ^target ^target reservoir ^target reservoir f0 Co where mtarget reservoir is the mass of the target substance inside the pixel reservoir, Ctarget_reservoir is the concentration of the target substance inside the reservoir, m0 is the mass of the target substance extracted through a single follicle, and Co is the concentration of the target substance extracted through a single follicle (into a reservoir of the same volume as the reservoir of interest), where m0 and Co can be determined based on the value for / 0■ In embodiments where voltage is detected instead of current, a corresponding quantisation method may be performed to quantise an amount of voltage Uo per dermal channel, and to determine a number of channels in a given reservoir as N = Utarget where Utarget is a measured voltage for the given reservoir. Again, this is related to mass as N = = m^et reserVoir = ctarget reservotr Uq mQ Cq The flux of the target substance jtarget can then be determined and compared to the reference flux ireference to estimate the subdermal concentration of the target substance Ctarget_subject, as discussed above in relation to steps 406 to 410 of Figure 4. Co-monitoring sensor Figure 10 shows an embodiment of a sensor 600 according to another embodiment of the invention. Similarly to the sensor 100 of Figure 1, the sensor 600 has a substrate 602 having a first layer 604 and a second layer 606. The sensor 600 has a plurality of sensor pixels 610, each of which comprises a reservoir and a set of electrodes (not shown), similarly to the sensor pixels 110 of the sensor 100. For brevity, these are not explained again here. The sensor 600 differs from the sensor 100 primarily in that the plurality of sensor pixels 610 of the sensor 600 includes two co-monitoring sensor pixels 614 as well as a plurality of additional sensor pixels 616. The co-monitoring sensor pixels 614 enable the reference flux jreference to be detected experimentally for a given subject, rather than using a pre-determined value. In contrast, the sensor 100 only included target-only sensor pixels, without any co-monitoring sensor pixels. In this embodiment, the additional sensor pixels 616 are configured as target-only sensor pixels, similarly to the pixels 110 in the sensor 100. In variant embodiments, the additional sensor pixels 616 may be configured as co-monitoring sensor pixels. The sensor 600 has six target-only sensor pixels, positioned in a 2x3 array. It will be appreciated that other numbers of target-only sensor pixels are also possible. Figure 11 shows one embodiment of a co-monitoring sensor pixel 614 which could be implemented in the sensor 600 of Figure 10. The co-monitoring sensor pixel 614 of Figure 11 is particularly suited to detecting substances that are primarily extracted by electro-osmosis rather than electro-migration, e.g. neutral (non-charged) substances (such as glucose and urea). The co-monitoring sensor pixel 614 includes a reservoir 608 embedded in a substrate portion (i.e. a portion of the substrate 602 of the sensor 600 in Figure 10). The co-monitoring sensor pixel 614 further includes a set of electrodes 612 aligned over the reservoir 608. Similarly to the sensor pixel 110, the reservoir 608 may be in the first layer 604 of the substrate and the electrodes 612 may be on or in the second layer 606 of the substrate 602. The set of electrodes 612 includes a first set of detection electrodes 618 and 620, a second set of detection electrodes 622 and 624, and an extraction electrode 626. The extraction electrode 626 is configured, together with an extraction electrode 626 of a neighbouring (adjacent) co-monitoring pixel 614, to apply an extraction current to the subject’s skin in use to cause transdermal iontophoretic extraction of the target and reference substances from the subject into the reservoir 608 (and into the reservoir of the neighbouring co-monitoring pixel 614). The first set of detection electrodes includes a first Working electrode (WE) 618 and a first combined Reference (RE)ZCounter (CE) electrode 620. The first set of detection electrodes is configured to detect the current Itarget that corresponds to the amount of the target substance extracted into the reservoir in use. The second set of detection electrodes includes a second WE 622 and a second RECE electrode 624. The second set of detection electrodes is configured to detect the current Inference that corresponds to the amount of the reference substance extracted into the reservoir in use. In this manner, the currents Itarget and Ireference can both be measured for the two substances through the same number of dermal channels, i.e. the number of dermal channels aligned with the reservoir 608. The reservoir 608 may contain a gel that is electrically conductive to help the first and second detection electrodes 618, 620, 622, and 624 detect their respective substances within the reservoir 608. Since the co-monitoring sensor pixel 614 has a relatively simple and compact sensor configuration, in some embodiments, it may be implemented in a majority or all the sensor pixels 610 of the sensor 600 (i.e. in a majority of or all the additional sensor pixels 616). Figure 12 shows an alternative embodiment of a co-monitoring sensor pixel 714, which could be implemented as the co-monitoring sensor pixel 614 in the sensor 600 of Figure 10 in other embodiments of the invention. The co-monitoring sensor pixel 714 is particularly suited to detecting charged reference substances (e.g. Na+). The co-monitoring pixel 714 has a round (e.g. disk-like or circular) footprint. At a centre of the co-monitoring pixel 714 (as viewed laterally along the pixel 714), there is provided an inner extraction electrode 725. In this embodiment, the inner extraction electrode 725 is circular. The inner extraction electrode 725 is operated as an anode. The co-monitoring pixel 714 includes a plurality (e.g. three) of sub-regions 715A-C having respective reservoirs 708A-C into which both the target and reference substances can be extracted and detected. Each sub-region 715A-C has a respective set of target detection electrodes, reference detection electrodes, and an outer extraction electrode. In this embodiment, each sub-region 715A-C is configured to detect a target substance of glucose and a reference substance of Na+. Glucose may be detected electrochemically, for example, by chronoamperometry, using a WE and a combined RE / CE electrode. Na+ may be detected by potentiometry that measures the electromotive force (emf) between all-solid state Na+ ion selective electrode (ISE Na+) and a reference (RE Na+) electrode. The ISE Na+ and RE Na+ material realization (i.e. the materials needed to realise an Na+ ion selective membrane, and its related reference electrode) may be based on published procedures. Thus, in this embodiment, in sub-region 715A, the reservoir 708A is overlaid with target detection electrodes which include a working electrode 718A and a first reference / counter electrode 720A for generating a primary electric signal having a current Itarget that corresponds to an amount of glucose extracted into the reservoir 708. The reservoir 708A is also overlaid with reference detection electrodes which, in this embodiment, include an ISE electrode 722A and Reference electrode 724A for generating a reference electric signal having a voltage Ureference that corresponds to the amount of Na+ extracted into the reservoir 708A in use. In variant embodiments, the reference detection electrodes and target detection electrodes may be configured differently, e.g. to detect different substances and / or to detect a current rather than a voltage. In the sub-region 715, an (outer) extraction electrode 726A (which is operated as a cathode in use) extends across the reservoir 708A and separates the target detection electrodes 718A, 720A from the reference detection electrodes 722A and 724A. The sub-regions 715B and 715C have corresponding configurations to the sub-region 715A, with corresponding reservoirs 708B-C and sets of electrodes 718B-C, 720B-C, 722B-C, 724B-C, and 726B-C. The sub-regions 715B and 715C have the same features as the sub-region 715A except as discussed otherwise herein. The sub-regions 715A-C together extend around a circumference of the co-monitoring pixel, i.e. around a periphery of the inner (central) extraction electrode 725. The inner extraction electrode 725 is separated from the sub-regions 715A-C by a gap, i.e. a portion of the pixel substrate which does not include any electrodes thereon. Each outer extraction electrode 726A-C is equidistant and symmetrically arranged around the inner extraction electrode 725, so that substantially the same electric field can be established within each sub-region 715A-C. The inner extraction electrode 725 (anode) may be aligned on top of a fourth reservoir portion 708D. The reservoir portion 708D under the inner extraction electrode 725 allows the extraction electrode 725 to act as an anode, whilst the reservoir portions 708A-C under the outer extraction electrodes 726A-C allow the outer extraction electrodes to act as cathodes. In this embodiment, the anode and cathodes are present in a single co-monitoring pixel. In other embodiments, the anodes and cathodes may be present in different (e.g. adjacent) pixels, e.g. with one pixel containing the anode and the other pixel containing the cathode. In this embodiment, the reservoirs 708A-C each have different surface areas, with reservoir 708A being the largest and reservoir 708C being the smallest. This allows the amount of current received through a given dermal channel to be separated out from the current received from skin without dermal channels . The reasoning for the different sizes of reservoirs 708A-C is explained below. With charged substances (e.g. Na+), the main mechanism driving extraction is electromigration, in contrast to neutral molecules (e.g. glucose, urea) which are extracted primarily via electroosmosis. Thus, extraction of charged substances takes place through the entire reservoir area, not only through dermal channels. However, in embodiments where the target substance is neutral (and thus only extracted through dermal channels), it is the extraction of the reference substance through the dermal channels that should be compared to the detected amount of the target substance, because then both the reference substance and target substance (e.g. Na+ and glucose) will have been extracted through the same pathways. In order to determine the flux of the reference substance through the dermal channels, one needs to separate the amount of the reference substance extracted via dermal channels from the amount extracted from skin without dermal channels (i.e. non-dermal channel contribution). Comparing extraction from sub-regions 715A-C having different reservoir areas allows the non-dermal channel contributions to be identified and separated from the dermal channel contributions. The core of this principle is shown in Figures 12 and 13. Figure 13 shows a schematic diagram of three sub-regions 715A-C each having reservoirs 708A-C of different surface areas Slt S2, and S3 respectively. In this example, each reservoir 708A-C overlies a single dermal channel (e.g. hair follicle) of cross-sectional area a. The total flux of the reference substance into the reservoir 708A, / re / erenee , including contributions from both dermal channels and other non-channel areas of the skin (i.e. areas of the skin without dermal channels, which may be referred to as non-dermal channels) can be represented according to equation (3a) below: Jreference_Sl ireference_non—dermal C$1 + jreference^l (3a) where jreference_non-dermaiis the non-channel flux (or “non-dermal channel” flux) of the reference substance per unit area (i.e. the flux through areas of the skin where there are no dermal channels, which may also be referred to as a “non-dermal channel flux”), is the area of the reservoir 708A, ireferenceis a flux of the reference substance through a single dermal channel (n = 1), and N1 is the number of dermal channels 204 contributing to the sub-region 715A (in this example, N = 1). Corresponding equations (3b) and (3c) can also be formulated for the remaining subregions of different areas S2 and S3, as: Jreference_S2 jreference_n<m—dermal ' ^2 ' ^2) T ireference ^2 (3b) Jreference_S3 ireferencenon—dermal ' C$3 ' ^3) T ireference (3c) Where Jreference_s2 and Jreference_s3 are the total fluxes into the reservoirs 715B and 715C respectively, and N2 and N3 are the number of dermal channels providing pathways into the reservoirs 715B and 715C respectively. Since a may be very small (e.g. corresponding to approximately 50 microns diameter for a hair follicle) compared to Slt S2, and S3 (which may be many mm2), the above equations (3a) to (3c) may be approximated as follows assuming a = 0: Jreference_Sl jreference_non—dermal ' "I" jreference^l Jreference_S2 jreference_non—dermal ' "I" jreference^2 Jreference_S3 jreference_non—dermal ' ^3 "I" jreference^3 The number of dermal channels Nlt N2, and N3 may be determined e.g. using the method previously described in relation to Figures 5 to 9. In order to solve ^or jreference non—dermal ^nd jreference ’ the total flUXeS Jreference_Sl ’ Jreference_S2 ’ ^nd Jreference_s3 are Arst determined. This can be calculated as ^reference Sl ' extract lonSl ^reference S2 ' ^S2 At ' extract i.on_S2 ^reference_S3 ' Ks3 ' Iextractlon_S3 Where Iextraction S1, I extractions^, and Iextraction S3 are the extraction currents applied to the respective reservoirs 708A-C; 1^, VS2, and are the volumes of the reservoirs 708A-C respectively; and Creferenee_si, Creferenee_s2, and Creference _S3 are the concentrations of the reference substance in the respective reservoirs 708A-C. The concentrations Creference S1, Creference S2, and Creference_S3 can be calculated based on a ratio of a respective measured reference voltages Ureference_si, ^reference_s2, and Ureference_S3 divided by the known sensitivity a for each respective sub-region 715A-C. The applied extraction currents may be selected such that an extraction current density through the reservoirs 708A-C is kept constant, i.e. such that: 1 extract lonSl ^extractlon S2 ^extractlon S3 Si S2 S3 The total fluxes ;re / erence S1,;re / erence S2, and ]reference_S3 may then be compared to determine the reference flux jreference through a given number of dermal channels n. For example, the total fluxes may be plotted against the reservoir areas as shown in Figure 14, and a line of best fit may be determined. The reference flux jreference Per dermal channel (n = 1) is estimated as the y-intercept of the best-fit line. The non-dermal channel flux of the reference substance per unit area jreferencenon-dermaiis estimated as the slope of the best-fit line. With these values being known, the reference flux inference Per dermal channel can then be used to estimate the subdermal concentration Ctargetsubject of the target substance in the usual manner. To ensure that the above equations (3a), (3b), and (3c) provide a solution, it can be advantageous to ensure that the equations will never reduce to being equivalent to each other. This can be achieved, e.g. by ensuring that the reservoir areas S2, and 53 of the sub-regions 715A-C are in a ratio of prime numbers. For example, returning to Figure 12, the reservoirs 708A-C have area ratios of 13:11:7. The circular arrangement of Figure 12 implements the principles shown in Figures 13 and 14 in a space-efficient manner, by positioning the reservoirs 708A-C around the inner (central) extraction electrode 725. This arrangement also creates an environment that is similar in terms of device layout for all three sub-regions 715A-C. However, other device configurations will also be possible according to other embodiments. In Figure 12, extraction measurements can be performed using the (outer) extraction electrodes 726A-C of the sub-regions 715A-C as cathodes and the (inner) extraction electrode 725 as a common anode. However, to increase the number of non-equivalent equations possible from equations (3a), (3b), and (3c), and thereby increase the amount of data available to evaluate the flux, extraction measurements can also be formed using different combinations of the (outer) extraction electrodes 726A-C as cathodes. Thus, measurements can be collected using the following combinations of extraction electrodes. Cathode Anode Extraction electrode 726A of first subregion (area S1) Common extraction electrode 725 (area S) Extraction electrode 726B of second subregion (area S2) Common extraction electrode 725 (area S) Extraction electrode 726C of third subregion (area S3) Common extraction electrode 725 (area S) Combination of extraction electrodes 726A and 726B (areas S1+S2) Common extraction electrode 725 (area S) Combination of extraction electrodes 726A and 726C (areas S1+S3) Common extraction electrode 725 (area S) Combination of extraction electrodes 726B and 726C (areas S2+S3) Common extraction electrode 725 (area S) This allows additional data points to be plotted (e.g. in the graph of Figure 14) using areas 51 + 52, 51 + 53, and 52 + 53. This corresponds to a scenario (in Figure 14) in which Nr = N2 = N3 = 1. In other scenarios where Nlt N2, and N3 are not equal, the values for jreference and area ireferencenon-dermai may stiII be solved for using the equations provided above. It will be appreciated that Figure 14 is shown to help illustrate the concepts described above, and that it is not essential for such a plot to be generated and displayed under embodiments of the invention. The co-monitoring sensor pixel of Figure 12 may be implemented as the co-monitoring sensor pixel 614 of sensor 600. In embodiments, only a minority (e.g. only one) of the sensor pixels in the sensor 600 may comprise the co-monitoring sensor pixel 714 that has a plurality of reservoirs of different areas. The co-monitoring sensor pixel 714 may be controlled (e.g. using a processing unit having a controller) to perform extraction less frequently than the remaining (target-only) sensor pixels, in order to avoid excessive accumulation of the reference substance (e.g. Na+) in the reservoirs 708A-C. Proportionality constant K As discussed above, the concentration of a target substance in a subject can be estimated based on equation Ctarget subject = .ltarget x Creference subject x K, where K is a predetermined proportionality Jreference constant between the reference substance and the target substance. The derivation of this equation is explained below. The measured target flux through a selected number of dermal channels n is proportional to the unknown concentration in the subject, i.e: j target ^t ' ^targetsubject (4) where Kt is a proportionality constant of transport through n dermal channels (e.g. one dermal channel) for the target substance. The value for Kt may be unknown as it may differ from person-to-person based on different characteristics of dermal channels (e.g. different dimensions of dermal channels). The reference substance may therefore be used to help account for any such variability. A similar relationship can thus be written for the reference substance, as: jreference ^r ' ^referencesubject (5) where Kr is a proportionality constant of transport through n dermal channels (e.g. one dermal channel) for the reference substance. Again, the value for Kr may be unknown. The values for Kt and Kr may depend on geometrical characteristics of the dermal channel (e.g. its cross-sectional area), which may be unknown. By using reference substance measurements and target substance measurements obtained from the same dermal channel(s), the dependency on these geometrical characteristics can be reduced or eliminated. By definition, the constants Kt and Kr can be combined to form a further constant K, as: Kr (6) K therefore takes into account the difference between the target substance’s transport process and the reference substance’s transport process. A value for K can be estimated theoretically or from dedicated prior experiments, for any combination of reference and target substances. Combining equations (4) and (6) then leads to: _ Kr j target ' ^target subject Further substituting in equation (5) for Kr leads to: _ Jreference 1 ~ ]target ' 77 ' '-‘targetsubject ^reference subject This equation can then be rearranged to derive the equation Ctargetsubject = .}target x Jreference Creference-subject x K, which can thus be used to determine the target concentration, in a manner that helps omit any variability introduced by the geometric characteristics of the dermal channels themselves (since these may be cancelled out by the ratio K = and by the reference and target substances being measured through the same dermal channels). The features disclosed in the foregoing description, or in the following claims, or in the accompanying drawings, expressed in their specific forms or in terms of a means for performing the disclosed function, or a method or process for obtaining the disclosed results, as appropriate, may, separately, or in any combination of such features, be utilised for realising the invention in diverse forms thereof. While the invention has been described in conjunction with the exemplary embodiments described above, many equivalent modifications and variations will be apparent to those skilled in the art when given this disclosure. Accordingly, the exemplary embodiments of the invention set forth above are considered to be illustrative and not limiting. Various changes to the described embodiments may be made without departing from the spirit and scope of the invention. For the avoidance of any doubt, any theoretical explanations provided herein are provided for the purposes of improving the understanding of a reader. The inventors do not wish to be bound by any of these theoretical explanations. Any section headings used herein are for organizational purposes only and are not to be construed as limiting the subject matter described. Throughout this specification, including the claims which follow, unless the context requires otherwise, the word “comprise” and “include”, and variations such as “comprises”, “comprising”, and “including” will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integers or steps. It must be noted that, as used in the specification and the appended claims, the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. Ranges may be expressed herein as from “about” one particular value, and / or to “about” another particular value. When such a range is expressed, another embodiment includes from the one particular value and / or 5 to the other particular value. Similarly, when values are expressed as approximations, by the use of the antecedent “about,” it will be understood that the particular value forms another embodiment. The term “about” in relation to a numerical value is optional and means for example + / -10%. The following numbered clauses, describing aspects of the invention, are part of the description: 1. A computer-implemented method for estimating a concentration of a target substance in a subject, the method comprising: receiving, a primary electric signal corresponding to an amount of the target substance that is extracted into a reservoir of a sensor through a set of dermal channels; determining a number of dermal channels N in the set of dermal channels; determining a flux of the target substance jtarget through a selected number of dermal channels n, based on the primary electric signal and the determined number of dermal channels W; obtaining a reference flux jreference of a reference substance through the selected number of dermal channels n at a reference concentration Creferencesubject of the reference substance; and estimating a concentration Ctarget subject of the target substance in the subject based on the flux of the target substance jtarget> reference flux jreference, and reference concentration r '~'reference_subject ■ 2. The computer-implemented method of clause 1, wherein determining the number of channels N comprises: receiving a plurality of electric signals each corresponding to an amount of a substance that is extracted into a respective reservoir of a plurality of reservoirs of the sensor; comparing the plurality of electric signals to quantise a value corresponding to an amount of the substance per dermal channel; and determining the number of channels for a given reservoir based on the electric signal corresponding to said reservoir and the quantised value. 3. The computer-implemented method of clause 1 or 2, wherein obtaining the reference flux jreference comprises selecting a pre-determined substantially constant value for flux of the reference substance at the reference concentration Creference_subjeCf 4. The computer-implemented method of clause 3, wherein the reference flux is selected based on one or more of the following characteristics of the subject: weight, skin follicle maturation cycle, body-mass index, race, age, sex, or medical condition. 5. The computer-implemented method of clause 3 or 4, wherein the reference substance is the same as the target substance. 6. The computer-implemented method of clause 1 or 2, wherein the reference substance is different from the target substance; wherein obtaining the reference flux jreference through the selected number of dermal channels n comprises: receiving a reference electric signal corresponding to an amount of the reference substance that is extracted into the reservoir of the sensor through the set of dermal channels; and determining the reference flux jreference through the selected number of dermal channels n, based on the reference electric signal, the determined number of dermal channels N, and the reference concentration of the reference substance Creference subject. 7. The computer-implemented method of clause 1 or 2, wherein the reference substance is different from the target substance; wherein obtaining the reference flux jreference through the selected number of dermal channels n, comprises: receiving three reference electric signals corresponding to amounts of the reference substance that are extracted into three reservoirs of the sensor respectively, wherein the three reservoirs each have different surface areas S2, and S3; determining a total flux of the reference substance in each reservoir ]reference_Si, Jreference_s2, Jreference_s3 using the reference electric signals; and comparing the total fluxes of the reference substance in each reservoir to determine the reference flux jreference through the selected number of dermal channels n. 8. A method for estimating a concentration of a target substance in a subject, the method comprising: generating the primary electric signal, using the sensor, by: using a set of extraction electrodes of the sensor to apply an extraction current to the subject’s skin to cause transdermal iontophoretic extraction of the target substance from the subject into the reservoir; and using a set of detection electrodes of the sensor to detect the primary electric signal that corresponds to the amount of the target substance extracted into the reservoir; wherein the method further includes performing the computer-implemented method of any preceding clause using the primary electric signal. 9. The method of clause 8, wherein the method comprises monitoring the concentration of the target substance over a set time period, by periodically repeating the steps for generating the primary electric signal, receiving the primary electric signal, determining the flux of the target substance jtarget’ and estimating the concentration of the target substance. 10. A method for estimating a concentration of a target substance in a subject, the method comprising: generating reference electric signal(s), using the sensor, by: using a set of extraction electrodes of the sensor to apply an extraction current to the subject’s skin to cause transdermal iontophoretic extraction of the reference substance from the subject into one or more reservoirs of the sensor; and using a set of detection electrodes of the sensor to detect one or more reference electric signal each corresponding to an amount of the reference substance extracted into a respective reservoir; and performing the computer-implemented method of clause 6 or 7 using the one or more reference electric signals to determine the reference flux jreference. 11. The method of clauses 9 and 10, wherein the reference electric signal(s) are generated less frequently within the time period than the primary electric signal. 12. An apparatus for extracting and detecting a target substance from a subject, the apparatus comprising: the sensor; and a processing unit having at least one processor and at least one memory including computer program code, wherein the computer program code is configured to, with the at least one processor, cause the processing unit to perform the computer-implemented method of any one of clauses 1 to 7. 13. A sensor for extracting and detecting a target substance from a subject, the sensor comprising: a sensor pixel assembly comprising a co-monitoring sensor pixel having: a substrate portion for positioning on the subject’s skin, the substrate portion comprising a reservoir assembly having one or more reservoirs; and a set of electrodes including: one or more extraction electrodes configured to apply a current to the subject’s skin to cause transdermal iontophoretic extraction of the target substance and a reference substance into each reservoir of the one or more reservoirs; one or more target detection electrodes configured to detect a primary electric signal that corresponds to an amount of the target substance extracted into each reservoir of the one or more reservoirs; and one or more reference detection electrodes configured to detect a reference electric signal that corresponds to an amount of the reference substance extracted into each reservoir of the one or more reservoirs. 14. The sensor of clause 13, wherein the reservoir assembly includes three reservoirs having different surface areas. 15. The sensor of clause 14, wherein the areas of the three reservoirs are non-divisible by each other, optionally wherein the areas of the three reservoirs are in a ratio of prime numbers. 16. The sensor of clause 14 or 15, wherein the three reservoirs are arranged around a periphery of at least one of the one or more extraction electrodes. 17. The sensor of any one of clauses 13 to 16, wherein the sensor pixel assembly comprises one or more additional sensor pixels, each additional sensor pixel having: an additional substrate portion for positioning on the subject’s skin, the additional substrate portion comprising an additional reservoir assembly having one or more additional reservoirs; and a set of additional electrodes including: an additional one or more extraction electrodes configured to apply a current to the subject’s skin to cause transdermal iontophoretic extraction of the target substance into each additional reservoir of the one or more additional reservoirs; and one or more additional target detection electrodes configured to detect a primary electric signal that corresponds to an amount of the target substance extracted into each additional reservoir of the one or more additional reservoirs. 18. The sensor of clause 17, wherein a majority of the sensor pixels are configured as co-monitoring sensor pixels. 19. The sensor of clause 17, wherein only a minority of the sensor pixels are configured as a comonitoring sensor pixel(s). 20. The sensor of any one of clauses 13 to 19, and a processing unit having at least one processor and at least one memory including computer program code, wherein the computer program code is configured to, with the at least one processor, cause the processing unit to perform the computer-implemented method of any one of clauses 1,2, 6, or 7.

Claims

1. A sensor for extracting and detecting a target substance from a subject, the sensor comprising: a sensor pixel assembly comprising a co-monitoring sensor pixel having:a substrate portion for positioning on the subject’s skin, the substrate portion comprising a reservoir assembly having one or more reservoirs; anda set of electrodes including:one or more extraction electrodes configured to apply a current to the subject’s skin to cause transdermal iontophoretic extraction of the target substance and a reference substance into each reservoir of the one or more reservoirs;one or more target detection electrodes configured to detect a primary electric signal that corresponds to an amount of the target substance extracted into each reservoir of the one or more reservoirs; andone or more reference detection electrodes configured to detect a reference electric signal that corresponds to an amount of the reference substance extracted into each reservoir of the one or more reservoirs.

2. The sensor of claim 1, wherein the reservoir assembly includes three reservoirs having different surface areas.

3. The sensor of claim 2, wherein the areas of the three reservoirs are non-divisible by each other, optionally wherein the areas of the three reservoirs are in a ratio of prime numbers.

4. The sensor of claim 2 or 3, wherein the three reservoirs are arranged around a periphery of at least one of the one or more extraction electrodes.

5. The sensor of any one of claims 1 to 4, wherein the sensor pixel assembly comprises one or more additional sensor pixels, each additional sensor pixel having:an additional substrate portion for positioning on the subject’s skin, the additional substrate portion comprising an additional reservoir assembly having one or more additional reservoirs; and a set of additional electrodes including:an additional one or more extraction electrodes configured to apply a current to the subject’s skin to cause transdermal iontophoretic extraction of the target substance into each additional reservoir of the one or more additional reservoirs; andone or more additional target detection electrodes configured to detect a primary electric signal that corresponds to an amount of the target substance extracted into each additional reservoir of the one or more additional reservoirs.

6. The sensor of claim 5, wherein a majority of the sensor pixels are co-monitoring sensor pixels.

7. The sensor of claim 5, wherein only a minority of the sensor pixels are co-monitoring sensor pixel(s).

8. An apparatus for extracting and detecting a target substance from a subject, the apparatus comprising:the sensor of any one of the preceding claims; anda processing unit having at least one processor and at least one memory including computer program code, wherein the computer program code is configured to, with the at least one processor, cause the processing unit to perform a computer-implemented method for estimating a concentration of a target substance in a subject, the method comprising:receiving a primary electric signal corresponding to an amount of the target substance that is extracted into a reservoir of the sensor through a set of dermal channels;determining a number of dermal channels N in the set of dermal channels;determining a flux of the target substance jtarget through a selected number of dermal channels n, based on the primary electric signal and the determined number of dermal channels N;obtaining a reference flux jreference of the reference substance through the selected number of dermal channels n at a reference concentration Creferencesubject of the reference substance; andestimating a concentration Ctarget subject of the target substance in the subject based on the flux of the target substance jtarget, reference flux jreference, and reference concentration r'~'reference_subject ■9. The apparatus of claim 8, wherein determining the number of channels N comprises:receiving a plurality of electric signals each corresponding to an amount of a substance that is extracted into a respective reservoir of a plurality of reservoirs of the sensor;comparing the plurality of electric signals to quantise a value corresponding to an amount of the substance per dermal channel; anddetermining the number of channels for a given reservoir based on the electric signal corresponding to said reservoir and the quantised value.

10. The apparatus of claim 8 or 9, wherein the reference substance is different from the target substance;wherein obtaining the reference flux jreference through the selected number of dermal channels n comprises:receiving a reference electric signal corresponding to an amount of the reference substance that is extracted into the reservoir of the sensor through the set of dermal channels; anddetermining the reference flux jreference through the selected number of dermal channels n, based on the reference electric signal, the determined number of dermal channels N, and the reference concentration of the reference substance Creference-subject ■11. The apparatus of claim 8 or 9, wherein the reference substance is different from the target substance;5 wherein obtaining the reference flux inference through the selected number of dermalchannels n, comprises:receiving three reference electric signals corresponding to amounts of the reference substance that are extracted into three reservoirs of the sensor respectively, wherein the three reservoirs each have different surface areas Sr, S2, and S3;10 determining a total flux of the reference substance in each reservoir Jreference_si,Jreference_s2, Jreference_s3 using the reference electric signals; andcomparing the total fluxes of the reference substance in each reservoir to determine the reference flux inference through the selected number of dermal channels n.

Images