SENSOR ARRAYS, METHOD FOR OPERATING A SENSOR ARRAY AND A COMPUTER PROGRAM FOR PERFORMING A METHOD FOR OPERATING A SENSOR ARRAY
20230389877 · 2023-12-07
Inventors
Cpc classification
International classification
A61B5/00
HUMAN NECESSITIES
Abstract
A sensor array comprises a base for providing a probe signal and a plurality of modular recording sites. Each modular recording site of the plurality of modular recording sites is configured for receiving a signal, for converting the signal into a digital sensor signal using an in-situ analog-to-digital converter and to provide the digital sensor signal to the base using a communication interface. The communication interfaces of the plurality of modular recording sites are connected serially with respect to each other and to the base and each in-situ analog-to-digital converter is configured for operating in a first operating mode and in a second operating mode. The base is configured for receiving a plurality of digital sensor signals from the plurality of modular recording sites and to process the plurality of digital sensor signals so as to provide the probe signal.
Claims
1. Sensor array comprising: a base for providing a probe signal; a plurality of modular recording sites, wherein each modular recording site of the plurality of modular recording sites comprises, a CMOS substrate, at least one sensor element configured for receiving an analog signal, an in-situ analog-to-digital converter configured for converting the analog signal into a digital sensor signal and a communication interface configured to provide the digital sensor signal to the base; wherein the communication interfaces of the plurality of modular recording sites are connected serially with respect to each other and to the base; wherein each in-situ analog-to-digital converter is configured for operating in a first operating mode for performing a first quantization of the analog signal using a first quantization setting and to acquire a residual error from the first quantization; and for operating in a second operating mode for performing a second quantization of the residual error using a second, different quantization setting for a same element of the analog-to-digital converter.
2. Sensor array of claim 1, wherein the in-situ analog-to-digital converter represents a continuous-time incremental delta-sigma converter.
3. Sensor array according to claim 1, wherein the first quantization represents a coarse quantization and the second quantization represents a fine quantization.
4. Sensor array according to claim 1, wherein each in-situ analog-to-digital converter comprises a signal input and is configured for providing a connection between the signal input and the signal in the first operating mode and for disconnecting the signal input from the signal in the second operating mode.
5. Sensor array according to claim 1, wherein the first operating mode and the second operating mode differ from each other in view of an amplification applied in a feedback loop of a delta-sigma modulator of the analog-to-digital converter; a sampling rate of the delta-sigma modulator; and/or of a signal shape applied for sampling in the delta-sigma-modulator.
6. Sensor array according to claim 1, wherein in the second operating mode an integrator and/or a quantizer and/or a feedback DAC of a delta-sigma modulator of an analog-to-digital-converter is reused with respect to the first operating mode.
7. Sensor array according to claim 1, wherein a delta-sigma-modulator of a digital-to-analog converter is implemented in absence of an input-feedforward path.
8. Sensor array according to claim 1, wherein each modular recording site of the plurality of modular recording sites comprises a communication interface configured to receive configuration data from the base, wherein the modular recording site is configured for adapting a parameter relating to the operation of the modular recording site based on the received configuration data.
9. Sensor array according to claim 1, wherein each modular recording site of the plurality of modular recording sites comprises a communication interface, wherein the communication interface comprises a serial interface, wherein the communication interfaces of the plurality of modular recording sites are connected to each other in a serial communication chain comprising a forward path from the base to a sensor array endpoint of the sensor array and a backward path from the sensor array endpoint to the base, wherein for each pair of a first modular recording site and a neighboring second modular recording site, the communication interface of the first modular recording site is connected to the forward path and the communication interface of the second modular recording site is connected to the backward path.
10. Sensor array according to claim 1, wherein a first subset of the plurality of modular recording sites is arranged on a first semiconductor substrate, and wherein an adjacent and neighboring second subset of the plurality of modular recording sites is arranged on a second semiconductor substrate; wherein the first semiconductor substrate and the second semiconductor substrate are spaced apart from each other by a gap and electrically connected to each other by at least one conductive line.
11. Sensor array according to claim 1, wherein the base comprises a wired output interface for providing the probe signal, wherein a number of channels of the wired output interface is independent of the number of modular recording sites and independent of a cross section of the plurality of modular recording sites in a plane perpendicular to an axial extension of the sensor array.
12. Sensor array according to claim 1, wherein the plurality of modular recording sites is arranged between the base and a sensor array endpoint of the sensor array, wherein the sensor array forms a needle.
13. Sensor array according to claim 1, wherein at least one modular recording site of the sensor array is configured for sampling the signal with at least a first and a second sensor element and for multiplexing outputs of the first and the second sensor element into the digital sensor signal.
14. Sensor array according to claim 1, further comprising, in each modular recording site, an offset compensation circuit configured for compensating an offset in the biosignal.
15. Sensor array according to claim 1, further comprising a data compression unit configured for reducing a data rate of the digital sensor signal acquired by a modular recording site of the plurality of modular recording sites; wherein the data compression unit is configured for determining a difference between a first digital sensor signal acquired by the modular recording site during a first instance of time and a second digital sensor signal acquired by the modular recording site during a second, later instance of time.
16. Sensor array according to claim 15, wherein the base comprises the data compression unit and wherein the data compression unit is configured to reduce a respective data rate of each digital sensor signal provided from the plurality of modular recording sites to the base.
17. Sensor array according to claim 15, wherein the modular recording site comprises the data compression unit.
18. Sensor array according to claim 13, wherein each modular recording site comprises a reduction element configured to reduce a word size of the digital sensor signal provided by the respective in-situ analog-to-digital converter; wherein, for each modular recording site, the respective reduction element is configured to reduce the word size of the respective digital sensor signal by omitting a predefined number of high-order bits.
19. Analog-to-digital converter comprising a continuous-time delta-sigma modulator configured for operating in a first operating mode for performing a first quantization of an analog signal using a first quantization setting and to acquire a residual error from the first quantization; and for operating in a second operating mode for performing a second quantization of the residual error using a second, different quantization setting of the analog-to-digital converter.
20. Neuronal probe comprising an analog-to-digital converter according to claim 19 and an offset compensation circuit, the compensation circuit configured for compensating an offset in the analog signal.
21. Method for operating a sensor array according to claim 1, comprising: recording of a signal with a sensor of a modular recording site of a plurality of modular recording sites of the sensor array; converting of the signal into a plurality of digital sensor signals using the plurality of modular recording sites of the sensor array by operating each analog-to-digital converter in the first operating mode and in the second operating mode to acquire the respective digital sensor signal; providing of the plurality of digital sensor signals to the base of the sensor array using the communication interfaces of the plurality of modular recording sites of the sensor array; receiving of the plurality of digital sensor signals from the plurality of modular recording sites of the sensor array with the base of the sensor array; processing of the plurality of digital sensor signals by the base of the sensor array so as to acquire a probe signal; and providing the probe signal with the base of the sensor array for a remote device.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0072] Embodiments of the present invention will be detailed subsequently referring to the appended drawings, in which:
[0073]
[0074]
[0075]
[0076]
[0077]
[0078]
[0079]
[0080]
[0081]
[0082]
[0083]
[0084]
[0085]
[0086]
[0087]
[0088]
[0089]
[0090]
[0091]
[0092]
[0093]
[0094]
[0095]
[0096]
[0097]
[0098]
[0099]
[0100]
[0101]
[0102]
[0103]
[0104]
[0105]
[0106]
[0107]
[0108]
[0109]
[0110]
[0111]
[0112]
[0113]
[0114]
[0115]
[0116]
[0117]
[0118]
DETAILED DESCRIPTION OF THE INVENTION
[0119] Equal or equivalent elements or elements with equal or equivalent functionality are denoted in the following description by equal or equivalent reference numerals even if occurring in different figures.
[0120] In the following description, a plurality of details is set forth to provide a more thorough explanation of embodiments of the present invention. However, it will be apparent to those skilled in the art that embodiments of the present invention may be practiced without these specific details. In other instances, well known structures and devices are shown in block diagram form rather than in detail in order to avoid obscuring embodiments of the present invention. In addition, features of the different embodiments described hereinafter may be combined with each other, unless specifically noted otherwise.
[0121] In the following, reference is made to embodiments of the present invention. Embodiments will be described in connection with neuronal probes as one possibility of implementing the present invention. Without any limitation, the description given hereinafter also relates to other sensor arrays, in particular biomedical sensor arrays. Examples for such biomedical sensor arrays are optical sensor arrays which may be used in connection with a retina of a human or animal. E.g., such a sensor array may be configured for receiving a signal. Although, in the following the described examples may refer to biosignals as the signal received by a modular recording site, the examples are not limited hereto but relate in general to other types of analogue signals such as, e.g., an optical signal or an electrical signal. A biosignal may be any signals in living beings that can be continually measured and monitored. The term biosignal may be used to refer to bioelectrical signals, but it may refer to both, electrical (e.g., electrochemically triggered) and non-electrical signals such as electrochemical signals or optical signals. A biosignal may in particular be or at least comprise an electrical signal, a signal being based on a biochemical reaction and/or an optical signal or stimulus. The sensor array may detect such a biosignal and may provide for a sensor signal based thereon. Thus, although the described examples may refer to neuronal probes which may be configured for receiving a neuronal signal, the examples are not limited hereto.
[0122]
[0123] The neuronal probe 100 may be used as a tissue penetrating probe for high density deep-brain recording of in vivo neural activity and overcomes a limitation by the level of electronic integration on the probe shank. Active probes are, in the prior art, used to improve the signal quality and reduce parasitic effects in situ, but still need to route these signals from the electrodes to a base where the readout electronics is located on a large area [4, 6]. The neuronal probe 100 comprises the conversion of the received neuronal signal 132.sub.1 to 132.sub.n into a digital sensor signal 134.sub.1 to 134.sub.n in each modular recording site 122.sub.e of the plurality 120 of modular recording sites, so that the base 110 does not need this component and therefore, the base 110 can be implemented on a small area.
[0124]
[0125] The neuronal probe 200 in
[0126]
[0127]
[0128]
[0129] The neuronal probe 300 of
[0130] In other words, a digital part 350 of each modular recording site 320.sub.1 to 320.sub.n of the plurality of modular recording sites comprises, for example, the communication interface 328.sub.1 to 328.sub.n which is connected to its neighboring nodes (for example the communication interfaces 328.sub.1 to 328.sub.n of each neighboring modular recording site) by a serial interface. In that way, the converted results (for example, the digital sensor signals) can be linked to the results of the neighboring nodes and can be carried to an external terminal (for example the base 330 or an external device connected to the base) of the overall system (for example, the neuronal probe 300). In that way, for example, the number of connections from one modular recording site 320.sub.1 to 320.sub.n of the plurality of modular recording sites to a neighboring modular recording site 320.sub.1 to 320.sub.n of the plurality of modular recording sites (the modular recording site 320.sub.1 to 320.sub.n can, for example, also be understood as a sensor channel) is kept as low as possible and no analog signal (for example the received neuronal signal by each modular recording site 320.sub.1 to 320.sub.n) susceptible to interferences is carried to the outside (for example to the base or to an external device connected to the base) or from the base 330 (for example the outside) into a modular recording site 320.sub.1 to 320.sub.n (for example also understood as a sensor node). Since all neuronal signals (this may be an electrode signal or a sensor signal) are digitized (this means, for example, that each neuronal signal received by a modular recording site 320.sub.1 to 320.sub.n of the plurality of modular recording sites 320.sub.1 to 320.sub.n is converted into a digital sensor signal) directly on-site (for example on each modular recording site 320.sub.1 to 320.sub.n) the neuronal probe 300 (respectively, the system) requires only a very small number of lines to the outside (this can, for example, include lines from the base 330 to an external device but it can also include, for example, lines from each modular recording site 320.sub.1 to 320.sub.n to the base 330). Thus, the neuronal probe 300 not only reduces the complexity but also the size of the base 330 and shank (for example, the plurality of modular recording sites 320.sub.1 to 320.sub.n with the tip 310) of the neuronal probe 300, because only a reduced number of lines from each modular recording site 320.sub.1 to 320.sub.n of the plurality of modular recording sites 320.sub.1 to 320.sub.n is needed to transfer a digital sensor signal to the base 330. The serial connection of the communication interfaces 328.sub.1 to 328.sub.n of the plurality of modular recording sites 320.sub.1 to 320.sub.n with respect to each other and to the base also enables the neuronal probe to contact an arbitrary number of modular recording sites 320.sub.1 to 320.sub.n (respectively, sensor nodes).
[0131] Each of the digital sensor signals from each of the modular recording sites 320.sub.1 to 320.sub.n can be transferred to the base 330 as a combined sensor signal. Each modular recording site 320.sub.1 to 320.sub.n, for example, provides the converted neuronal signal as a digital sensor signal to the bi-directional serial digital data bus 328. Then the digital sensor signal is transferred by the digital data bus 328 as a combined sensor signal, of all digital sensor signals of the plurality of modular recording sites, to the base 330, where the base 330 processes the combined sensor signal to provide a probe signal.
[0132] According to an embodiment, the neuronal probe 100, 200 and 300 shown in
[0133] The sensor array 300 is according to an embodiment a Pixel-Level ADC (ADC=analog-to-digital converter), which represents, for example, an optical sensor with a conversion of an analog signal to a digital signal at each modular recording site 320.sub.1 to 320.sub.n. The inventive sensor array 300 is optimized with regard to known optical sensors in terms of a self-sufficient analog-to-digital-conversion. Known sensors are based on an electric current (photodiode) to time conversion (time interval of a pulse). The time interval of a pulse respectively the time needed to reach a threshold is measured in known sensors at all sensor knots or nodes at the same time, which results in a parallel link between the sensors and no serial link is possible. A serial link like in an embodiment of the sensor array 300 will result e.g. in area-efficient and/or cheaper sensors.
[0134] The base 330 of the neuronal probe 300 may comprises a reference 331 which can, for example, have the same functionality as the reference 212 of
[0135] Each converter (modular recording site 320.sub.1 to 320.sub.n) comprises a digital part 350 and an analog part 360 and is configured for minimum area consumption. The system (for example, the neuronal probe 300) is structured in a modular manner: the signal of each electrode (sensor element 322.sub.1 to 322.sub.n) is locally converted to a digital output signal independent of the neighboring electrode (sensor element 322.sub.1 to 322.sub.n).
[0136] In that way, the number of connections to a neighboring module (sensor channel/modular recording site 320.sub.1 to 320.sub.n) is kept as low as possible and no analog signals susceptible to interference are carried to the outside or from the outside into the sensor node (sensor element 322.sub.n to 322.sub.n). Since all electrode signals (neuronal signals) are digitized (converted into a digital signal) directly on site, the system (for example, the neuronal probe 300) requires only a very small number of lines to the outside (for example, to an external device).
[0137] Each modular recording site 320.sub.1 to 320.sub.n of the neuronal probe 300 can receive a neuronal signal with sensor elements 322.sub.1 to 322.sub.n and convert the received neuronal signal into a digital sensor signal by first integrating the neuronal signal by the integrator 324.sub.1 to 324.sub.n and quantization using a quantizer 325.sub.1 to 325.sub.n. With the bi-directional serial digital data bus 328, the digital sensor signal of each modular recording site 320.sub.1 to 320.sub.n is transferred from each modular recording site 320.sub.1 to 320.sub.n to the digital interface/control unit 339 of the base 330, where the digital sensor signal is processed by the digital interface/control unit 339 into a probe signal, wherein the probe signal is transferred over the pads/digital four wire interface 340 to an external device.
[0138] The whole probe (neuronal probe 300) is separated along its length into a digital 350 and an analog part 360 with separate supply routing and optionally a low-impedance ground shield 390 in between, that also covers the top to increase the robustness against EMI (electromagnetic interferences) and to reduce digital noise coupling.
[0139] The supply routing can be realized with the electrical power supply 334. The supply routing for the analog part 360 comprises, for example, a voltage V.sub.DD,A 335 with an optional supply buffer 335.sub.a and a ground voltage V.sub.SS 336. The supply routing for the digital part comprises, for example, a voltage V.sub.DD,D 337 with the ground voltage V.sub.SS 336. The digital part 350 may be connected to its neighboring node (neighboring recording site 320.sub.1 to 320.sub.n) by a serial interface (communication interfaces 328.sub.1 to 328.sub.n). In that way, the converted results may be linked to the results of the neighboring node (neighboring modular recording site 320.sub.1 to 320.sub.n) and may be carried to the external terminal (for example an external device or the base 330) to the overall system. Configuration data can be carried in a linked manner to each sensor node (modular recording site 320.sub.1 to 320.sub.n) via the same interface (communication interfaces 328.sub.1 to 328.sub.n) for switching the node (for example, the modular recording site 320.sub.1 to 320.sub.n) on and off or for changing its scaling, for example. The base 330 includes, for example, a reference transistor 332 providing the global voltage biasing V.sub.BIAS to all recording sites (modular recording sites 320.sub.1 to 320.sub.n) and a finite-state-machine that allows to switch between a configuration mode and a normal operating mode (readout modes) and forwards the internal data (for example the digital sensor signal) and configuration chains to the external unit (for example, an external device). The switching (of the communication interfaces 328.sub.1 to 328.sub.n) between the configuration mode 328.sub.1a to 328.sub.na and the normal operation mode 328.sub.1b to 328.sub.nb is controlled, for example, with a separate control signal. In the configuration mode 328.sub.1a to 328.sub.na, settings may be readout of a storage, which is, for example, transferred over the bi-directional serial digital data bus 328 and can be adapted by each modular recording site 320.sub.1 to 320.sub.n, while in the normal operation mode 328.sub.1b to 328.sub.nb the digital output (the digital sensor signal) of the converter (modular recording site 320.sub.1 to 320.sub.n) will, for example, be written in the storage. The storage may be a volatile or non-volatile storage and may comprise, for example, a plurality of transistor elements for storing information. According to embodiments, the configuration data in the configuration mode 328.sub.1a to 328.sub.na and the digital data in the normal operation mode 328.sub.1b to 328.sub.nb may be transmitted between each modular recording site 320.sub.1 to 320.sub.n and the base 330 without using a storage by using the bi-directional serial digital data bus 328 directly.
[0140] On the digital side 350, no global signals have to be routed: the chain signal, as well as the clock, may be forwarded from one block (bi-directional serial digital data bus 328) to the next one. The clock is slightly delayed from each modular recording site 320.sub.1 to 320.sub.n to each modular recording site 320.sub.1 to 320.sub.n to spread digital supply noise and reduce peak current consumption. The recording sites (modular recording sites 320.sub.1 to 320.sub.n) are grouped into blocks of two ADCs (analog-to-digital converter), one connected to the forward chain 352 and one to the backward chain 354 and clock. In the analog part 360, there may be only two global reference lines throughout the probe, i.e., the body reference voltage V.sub.BODY 333a and V.sub.BIAS 333b. The bias voltage (referenced to V.sub.DD) is routed with large parasitic capacitances to the supply 334 to enhance noise rejection from external sources.
[0141] In other words, the modular recording sites 320.sub.1 to 320.sub.n are, for example, grouped into blocks of two modular recording sites (ADCs respectively, analog-to-digital converters comprising, for example, each of the integrators 324.sub.1 to 324.sub.n and each of the quantizer 325.sub.1 to 325.sub.n), wherein one modular recording site comprising a first ADC is connected to the forward chain 352 and forward clock and the other modular recording site comprising a second ADC is connected to the backward chain 354 and backward clock. Thus, the serial interfaces (communication interfaces 328.sub.1 to 328.sub.n) of every second modular recording site are, for example, connected/coupled to a first chain (for example forward chain 352) and the serial interfaces of all other modular recording sites are connected/coupled to a second chain (for example backward chain 354). The first chain and the second chain are coupled to the base such that the digital sensor signal is transferred to the base.
[0142] The analog-digital converter (for example, the integrator 324.sub.1 to 324.sub.n coupled with the quantizer 325.sub.1 to 325.sub.n) may be configured with a differential input in order to be largely robust against interferences of the supply voltage (for example, the supply voltage V.sub.DD, 335 with the ground V.sub.SS 336 of the electrical power supply 334). Apart from the supply voltage (for example, the electrical power supply 334), two further global lines (for, example, joints) may be used and may be shared by all sensor nodes (modular recording sites 320.sub.1 to 320.sub.n): [0143] The control voltage (for example, the reference voltage V.sub.BIAS 333b): each circuit requires, for example, internally a certain number of constant reference potentials and setting currents. All currents and potentials are, for example, derived from the global adjusting voltage (for example, the control voltage V.sub.BIAS 333b) which is distributed to all sensor nodes (modular recording sites 320.sub.1 to 320.sub.n) as global line. Since it is a global line to which all sensor modules (modular recording sites 320.sub.1 to 320.sub.n) are connected, the same may be provided with large parasitic capacitance. This has a positive effect on possible coupling of noise via this line. Additionally, noise (for example, on the control voltage V.sub.BIAS 333b) is suppressed by the differential readout principle. This control voltage (for example, the reference voltage V.sub.BIAS 333b) may be used for testing, but in principle it (for example, the reference voltage V.sub.BIAS 333b) does not necessarily have to be a global connection and the function can, for example, be implemented under each electrode (sensor element 322.sub.1 to 322.sub.n). [0144] The reference voltage (for example, the reference voltage V.sub.BODY 333a): One side of the differential input is, for example, connected to the sensor (electrode/sensor element 322.sub.1 to 322.sub.n) while the second input is, for example, connected to a reference voltage (for example, the reference voltage V.sub.BODY 333a). Possible interferences on V.sub.BODY (reference voltage V.sub.BODY 333a) can be detected equally on all sensor nodes (modular recording sites) and can hence be filtered out in digital post-processing.
[0145] Since even the largest neuronal signals are only in the range of some tens of millivolts and the required linearity is low, a direct conversion using, for example, a continuous-time gm-C based incremental delta-sigma analog-to-digital converter (for example, the integrator 324.sub.1 to 324.sub.n coupled with the quantizer 325.sub.1 to 325.sub.n) under each electrode (sensor element 322.sub.1 to 322.sub.n) may be implemented using a first-order modulator allowing the implementation on a minimal silicon area, since only one integrator 324.sub.1 to 324.sub.n and capacitor and no accurate time constants, thus no local biasing, are needed. Decimation may be accomplished using a simple ripple counter 326.sub.1 to 326.sub.n. The output of the single branch OTA-C integrator 324.sub.1 to 324.sub.n is, for example, connected to the quantizer 325.sub.1 to 325.sub.n, i.e., comparator and output latch, driving the switches for the current feedback.
[0146] In the following a concrete example of a neuronal probe according to an embodiment is given, such as the neuronal probe 300. The example includes several concrete values of parameters used for implementing the neuronal probe. The values are to be understood as non-limiting example only and therefore they do not limit the embodiments but are merely suitable for a better understanding of the present invention. It is clear that by using further, different or other components other values may be obtained such as different voltages, currents and/or data rates.
[0147] The digital part 350 of the ADC (analog-to-digital converter) consists or comprises, for example, of a decimator, i.e., ripple counter 326.sub.1 to 326.sub.n, two registers for the 11b (11 bit) conversion result and a 2b (2 bit) configuration register. According to an example the ADC may run for 1024 cycles delivering a 10b result. Before resetting the OTA, the OTA output (for example, the output of the integrator 324.sub.1 to 324.sub.n) and the counter 326.sub.1 to 326.sub.n, the results may be transmitted using the data bus. For example, the last result of the comparator which represents the final conversion error may be appended as the eleventh bit to the 10 bit result and the obtained 11 bit may be put on the data chain. For example, the bit sequence may be stored in the storage of the bi-direction serial digital data bus 328. Alternatively or in addition, the bit sequence of the digital sensor signals of each modular recording site 320.sub.1 to 320.sub.n may directly be transmitted using the bi-directional serial digital data bus 328. The delayed clock of a following cell (modular recording site 320.sub.1 to 320.sub.n) is used for a latch to avoid timing violations between the latch and the comparator. During readout, the digital data (the plurality of digital sensor signals) is shifted through the modular recording sites 320.sub.1 to 320.sub.n and the neuronal probe 300 uses two possibly chains (the forward chain 352 and the backward chain 354), each of them may use the same or different data rate such as at least 15 Mbit/s or at least 20 Mbit/s such as 20.48 Mbit/s. For example both chains (the forward chain 352 and the backward chain 354) may use a bit rate of 20.48 Mbit/s, i.e., f.sub.s=20.48 MHz. The FSM in the base 330 may combine the outputs of both chains into a single data stream, e. g., by time multiplexing which yields, for example, at the front-end in the given example. The base 330 may be a low power element and may consume less than 1 W, less than 100 mW or even less than 100 μW, e. g., 37 μW and the power consumption per recording site (modular recording site 320.sub.1 to 320.sub.n) may result to less than 1 W, less than 100 mW or even less than 100 μW, e. g., 39.14 μW, of which less than 1 W, less than 100 mW or even less than 100 μW, e. g., 12.77 μW are consumed by the analog part 360.
[0148] The 11b ADCs (for example, the integrator 324.sub.1 to 324.sub.n coupled with the quantizer 325.sub.1 to 325.sub.n) may be designed to optimize noise performance per area, therefore as much area as possible is dedicated to the noise critical components, i.e., the input (the dedicated area should be less than 1000 μm.sup.2, or less than 500 μm.sup.2, or less than 200 μm.sup.2 such as 171 μm.sup.2 in the given example) and the load transistors (the dedicated area should be less than 1000 μm.sup.2, or less than 500 μm.sup.2, or less than 200 μm.sup.2 such as 144 μm.sup.2 in the given example). Only a small area is, for example, dedicated to the feedback current sinks, which are derived from the global bias line. The feedback current determines the full scale (FS) of the ADC (for example, the integrator 324.sub.1 to 324.sub.n coupled with the quantizer 325.sub.1 to 325.sub.n) which can be configured to one of three different scopes, for example, one scope of ±15 mV, a second scope of ±30 mV and a third scope of ±55 mV or one scope of ±13 mV, a second scope of ±25 mV and a third scope of ±50 mV, or in one scope of ±12 mV, a second scope of ±23 mV and a third scope of ±46 mV, such as ±11.25 mV, ±22.5 mV or ±45 mV.
[0149]
[0150] Each modular recording site 320.sub.1 to 320.sub.n may be divided into an analog part 360 and a digital part 350. The digital part 350 may be shielded from the analog part 360 by a first conductive element 390, wherein the conductive element is configured to block electromagnetic interferences. The conductive element can, for example, be a low-impedance ground shield. A second conductive element 392 is arranged encircling a connector (electrode contact 323) of the sensor element 322.sub.1 to 322.sub.n so as to shield the analog part 360 and the digital part 350 from the sensor element 322.sub.1 to 322.sub.n. Under each sensor element 322.sub.1 to 322.sub.n the reference 331 can be implemented with the control voltage V.sub.BIAS 333b and the reference voltage V.sub.BODY 333a. The electrical power supply 334 can be implemented as, for example, the analog supply voltage V.sub.DD,A 335 and the ground voltage V.sub.SS 336.
[0151] At least a part of the analog part 360 and at least a part of the digital part 350 are, for example, covered by a sensor element 322.sub.1 to 322.sub.n configured for receiving the neuronal signal. In an embodiment the sensor element 322.sub.1 to 322.sub.n covers the digital part 350 and the analog part 360 centric, but it is also possible, that the sensor element 322.sub.1 to 322.sub.n covers only the digital part 350 or only the analog part 360.
[0152] In this embodiment, a modular and scalable architecture of a needle probe is realized, which, instead of routing or pre-buffering noise-sensitive analog signals along the shank, integrates, for example, analog-to-digital conversion under each electrode (sensor element 322.sub.1 to 322.sub.n). An area used for such integration may be arbitrary and be influenced by the functionality to be integrated. For example, an area of less than 200×200 μm.sup.2, less than 150×150 μm.sup.2 or less than 100×100 μm.sup.2, such as, for example, 70×70 μm.sup.2 may be used. The design eliminates the need for any additional readout circuitry at the top of the probe (for example, neuronal probe 300) and connects with a digital 4-wire interface 340. The neuronal probe 300 may be implemented as reconfigurable 11.5 mm neuronal probe (but it is also possible to have larger probes of more than 11.5 mm, more than 14 mm or also more than 20 mm) features a constant width 370 of 70 μm (The width 370 is, for example, smaller than 100 μm, smaller than 90 μm or smaller than 75 μm) and thickness 380 (The thickness 380 is, for example, smaller than 100 μm, smaller than 80 μm or smaller than 60 μm) of 50 μm from top to bottom (for example, from the last modular recording site 320.sub.n to the first modular recording site 320.sub.1 or to the base 330 along an axial extension of the array) for minimal tissue damage with, for example, 144 integrated recording sites (modular recording sites 320.sub.1 to 320.sub.n) and can be fully immersed in tissue for deep-brain recording applications.
[0153] The plurality of modular recording sites 320.sub.1 to 320.sub.n is arranged along an axial direction and forms an array along the axial direction. An extension along a first perpendicular direction perpendicular to the axial direction is understood as the width 370 and an extension along a second perpendicular direction perpendicular to the axial direction understood as the thickness 380.
[0154] In an embodiment, the plurality of modular recording sites 320.sub.1 to 320.sub.n is, for example, arranged along an axial direction and forms an array along the axial direction. An extension of the base 330 along a first perpendicular direction perpendicular to the axial direction is, for example, at most an extension of the plurality of modular recording sites 320.sub.1 to 320.sub.n along the first perpendicular direction. An extension of the base 330 along a second perpendicular direction perpendicular to the axial direction is, for example, at most an extension of the plurality of modular recording sites along the second perpendicular direction.
[0155] In other words, an elongation of the base 330 in a direction of an axis of the plurality of modular recording sites 320.sub.1 to 320.sub.n may be not bigger than an elongation of the plurality of modular recording sites 320.sub.1 to 320.sub.n in a direction of an axis of the plurality of modular recording sites 320.sub.1 to 320.sub.n. An elongation of the base 330 in a direction perpendicular to an axis of the plurality of modular recording sites 320.sub.1 to 320.sub.n may be not bigger than an elongation of the plurality of modular recording sites 320, to 320.sub.n in a direction perpendicular to an axis of the plurality of modular recording sites 320.sub.1 to 320.sub.n.
[0156] In other words, the cross-section perpendicular to an axis from the base through all modular recording sites 320.sub.1 to 320.sub.n (through the plurality of modular recording sites 320.sub.1 to 320.sub.n) to the last modular recording site 320.sub.n does not have to change. This has the advantage that one can choose an arbitrary number of modular recording sites 320.sub.1 to 320.sub.n for the plurality of modular recording sites 320.sub.1 to 320.sub.n of the neuronal probe without influencing the cross-section of the base 330. Thus, the base 330 can, for example, have the same cross-section as each of the modular recording sites 320.sub.1 to 320.sub.n of the plurality of modular recording sites 320.sub.1 to 320.sub.n. Thus, it is possible to bury/immerse the base 330 completely in the tissue. Thus, the neuronal probe 300 can be placed deeper into the tissue and the invasive surgical procedure may be minimized.
[0157] The modular concept allows the realization of any arrangement of sensor nodes (modular recording sites 320, to 320.sub.n), such as, for example, in the form of a two-dimensional array of a needle having one or multiple columns. Each modular recording site 320, to 320.sub.n may be arranged and connected to each other in at least one row or column. When referring again to
[0158] Since only the digital data and no sensitive signals are carrier along the shank, a low amount or even no crosstalk can be measured between the sensors (for example, between each sensor element 322.sub.1 to 322.sub.n) or between each modular recording sites 320.sub.1 to 320.sub.n and there is high robustness with respect to external interference sources such as light sources or electromagnetic fields.
[0159] The number of lines to the outside may be independent of the number of electrodes (sensor elements 322.sub.1 to 322.sub.n) and/or sensor modules (for example, the number of modular recording sites 320.sub.1 to 320.sub.n), respectively, and/or independent on the width 370 of the shank.
[0160] Since the base 330 of the needle may have the same width 370 as the shank, the same can be introduced into tissue beyond the length of the shank without causing additional damage. Thus, deeper brain areas can be measured when compared with conventional needles.
[0161] The neuronal probe 300 can, for example, represent a fully immersible deep-brain neuronal probe with modular architecture and a Delta-Sigma ADC (analog-to-digital converter) integrated under each electrode for parallel readout of, for example, 144 recording sites.
[0162]
[0163] According to an embodiment the segmented probe 300 shown in
[0164] According to an embodiment, the sensor array, i.e. the segmented probe 300, comprises a plurality of modular recording sites, for example, represented by the segments 321.sub.1 to 321.sub.3 or by the sensor elements 322.sub.1 to 322.sub.30. For example, one sensor element 322.sub.1 to 322.sub.30 can represent one modular recording site and/or a group of two or more sensor elements 322.sub.1 to 322.sub.30 can represent one modular recording site, wherein one modular recording site comprises, for example, one analog-to-digital-converter.
[0165] According to an embodiment, a first subset of the plurality of modular recording sites is arranged on a first semiconductor substrate 329.sub.1, and wherein an adjacent and neighboring second subset of the plurality of modular recording sites is arranged on a second semiconductor substrate 329.sub.2; wherein the first semiconductor substrate 329.sub.1 and the second semiconductor substrate 329.sub.2 are spaced apart from each other by a gap and electrically connected to each other by at least one conductive line 386.sub.1 to 386.sub.6. If the first segment 321.sub.1 represents one modular recording site, this means, for example, that the segment 321.sub.1 is the first subset of the plurality of modular recording sites, wherein the modular recording site comprises 12 sensor portions 322.sub.1 to 322.sub.12 and one analog-to-digital-converter. If a sensor element 322.sub.13 to 322.sub.24 or groups of sensor elements (e.g. 2×2, 3×1, 1×3, 3×2, etc.) represent on the second semiconductor substrate 329.sub.2 two or more modular recording sites, this means, for example, that the two or more modular recording sites represent the second subset. Thus, for example, the second subset on the second semiconductor substrate 329.sub.2 comprises 4 modular recording sites, if each modular recording site comprises 3 sensor elements 322.sub.13 to 322.sub.15, 322.sub.16 to 322.sub.18, 322.sub.19 to 322.sub.21 and 322.sub.22 to 322.sub.24 with, for example, a 3×1 configuration and wherein each of the 4 modular recording sites comprises one analog-to-digital-converter. In
[0166] According to an embodiment, the at least one conductive line 386.sub.1 to 386.sub.6 is arranged on or in a flexible substrate 385.
[0167] According to an embodiment, the segments 321.sub.1 to 321.sub.3 are arranged on the flexible substrate 385, like a flexible polymer cable. To improve the flexibility of the segmented probe 300 and to reduce possible damage to the segmented probe 300 by bending or twisting the probe, a gap between the segments 321.sub.1 to 321.sub.3 can be adjusted. The larger the gap, the more flexible is, for example, the segmented probe 300.
[0168] According to an embodiment a segment 321.sub.1 to 321.sub.3 comprises contacts, which connect each segment 321.sub.1 to 321.sub.3 to the flexible polymer cable 385, the flexible polymer cable thereby interconnecting the segments, e.g., in a serial way. According to an embodiment signals and/or a power supply is provided for each segment 321.sub.1 to 321.sub.3 by the base through the flexible polymer cable 385 and each segment 321.sub.1 to 321.sub.3 is, for example, configured to transmit signals by the flexible polymer cable 385 to the base, wherein the segments 321.sub.1 to 321.sub.3 are, for example, connected serially by the flexible polymer cable 385. Each segment 321.sub.1 to 321.sub.3 comprises, for example, a number of interfaces or contacts 328.sub.1, 328.sub.2, 331, 335, 336 and/or 338. Whilst being illustrated as having six contacts, embodiments are not limited hereto as any number of contacts, i.e., more or less, may be implemented so as to transmit signals and/or provide for power supply. For example, a first contact 331 represents a reference, e.g. a reference voltage, a second contact 338 represents a clock, a third contact 328.sub.1 represents a data input, a fourth contact 328.sub.2 represents a data output, a fifth contact 335 represents a voltage supply and a sixth contact 336 represents a ground, e.g. a ground voltage.
[0169] According to an embodiment, the segments 321.sub.1 to 321.sub.3 can comprise a CMOS silicon substrate 329.sub.1 to 329.sub.3, in which the electronics of the proposed sensor array, i.e. the segmented probe, is implemented. Optionally the six contacts or at least some of the six contacts 331, 338, 328.sub.1, 328.sub.2, 335 and/or 336 can be realized as through-silicon vias.
[0170]
[0171] The quantizer 430 can, for example, have the same functionality as each quantizer 325.sub.1 to 325.sub.n of
[0172] The sensor element 410 of the modular recording site 400 can, for example, receive a neuronal signal and transfer the neuronal signal to the continuous-time Gm-C Delta-Sigma modulator 420. The continuous-time Gm-C Delta-Sigma modulator 420 can then, for example, integrate the neuronal signal and send the integrated neuronal signal to the quantizer 430, wherein the quantizer 430 can, for example, decimate the integrated neuronal signal and convert the neuronal signal into a digital sensor signal. The control unit 440 is, for example, configured to be either in a normal operating mode or in a configuration mode. The control signal 442 may tell the control unit 440 which mode is appropriate. When the control unit operates in the normal operating mode, the digitized sensor signal from the quantizer 430 may be written in the storage 460 and transferred by a digital data bus 462 to a base. When the control unit 440 operates in the configuration mode, the control unit 440 may transmit configuration parameters to the configuration module 450, wherein the configuration module 450 can thereby, for example, change parameters for operating each modular recording site such as, for example, the scaling of the continuous-time Gm-C Delta-Sigma modulator 420 is changed.
[0173] In other words, the communication interface (for example, control unit 440 and the storage 460) can, for example, operate either in a configuration mode or in a normal operating mode. The switching between the configuration mode and the normal operating mode can, for example, be carried out by a separate control signal. In the configuration mode, the setting of each modular recording site can, for example, be read out of the received configuration data and be used by each modular recording site for adapting a parameter relating to the operation of the modular recording site 400. In a normal operating mode of the communication interface of each modular recording site of the plurality of modular recording sites, the digital sensor signals can be transferred to the base. With the implementation of the communication interface in each modular recording site it is possible to operate each modular recording site individually and to change parameters regarding the conversion of the neuronal signal into a digital sensor signal.
[0174] For example, a digital data bus 462 comprising a storage 460 (the storage 460 is, for example, a transistor) couples the communication interface of each modular recording site of the plurality of modular recording sites serially with respect to each other and to the base. The line for the control signal 442 and the line for the digital data bus 462 do not necessarily have to be individual lines. It is, for example, possible to transmit the control signal 442 using the data bus 462. If the communication interface is in a normal operating mode the digital sensor signal converted by each modular recording site of the plurality of modular recording sites from the neuronal signal can, for example, be written in the storage 460 of the digital data bus 462. The digital data bus 462 then, for example, transfers the storage with the digital sensor signal to the base. The communication interfaces of each modular recording site of the plurality of modular recording sites are connected serially with respect to each other and to the base. This can mean that each communication interface of each modular recording site writes the respective digital sensor signal in the storage 460 of the digital data bus 462, so that the digital sensor signals of each modular recording site are arranged in sequence respective to a position of each modular recording site with respect to a position of each other modular recording site. With a separate control signal, the communication interface can change its mode from a normal operation mode (where the digital sensor signal is transferred from each modular recording site to the base) to configuration mode (where configuration data is transferred from the base to each modular recording site). Thus, this modular system concept allows, for example, the contacting of an arbitrary number of modular recording sites (for example neuronal electrodes of any topology or geometry) with minimum complexity, a small size of the base and reading out the modular recording sites simultaneously.
[0175] In other words, the architecture of an analog-to-digital converter (respectively, the modular recording site 400) resembles the architecture of a time-continuous Gm-C Delta-Sigma modulator. The time-continuous Delta-Sigma modulators are known for their reduced demand of electrical power. An implementation with a Gm-C integrator 420 has additionally the advantage that the area demand can be very small. Furthermore, these converters are known that they comprise an implicit anti-aliasing filter effect. Thus, more electrical power and area can be saved, because the necessity for a dedicated anti-aliasing filter as an additionally circuit block can be omitted. The usage of such converters and circuit architecture makes it possible to reduce the signal chain and to put an analog-to-digital converter directly under a sensor element 410.
[0176]
[0177] The at least two sensor elements 410.sub.1 and 410.sub.2 are adapted for receiving a signal. The two or more sensor elements 410.sub.1 to 410.sub.4 are, for example, arranged in a 2×2-sensormatrix, representing four sensor elements 410.sub.1 to 410.sub.4. According to an embodiment all sensor elements 410.sub.1 to 410.sub.4 of one modular recording site 400 can detect the same signal, whereby each sensor element 410.sub.1 to 410.sub.4 generates for example a different individual signal associated with the same signal, based on a dependency of the same signal on different positions of the sensor elements 410.sub.1 to 410.sub.4 at the modular recording site 400. That is, the same signal, the biosignal for example, is received at the different sensor elements 410.sub.1 to 410.sub.4 such that the sensor elements 410.sub.1 to 410.sub.4 provide for individual, different signal being based on the same biosignal as described in connection with other embodiments of the present invention.
[0178] Thus, according to an embodiment, the modular recording site 400 is configured to process four individual signals, associated with the received signal. The individual signals can differ from each other or at least some of the individual signals can be the same. The integrator 420 may integrate the individual signals of the at least to sensor elements 410.sub.1 and 410.sub.2 in a time-sequential manner, thereby allowing the modular recording site 400 for sequentially providing an output signal based on two or more sensor elements, i.e., to time-multiplex the information collected with the sensor elements 410.sub.1 and 410.sub.2. This may allow for a higher resolution and/or for an oversampling of the signal sensed with the sensor elements 410.sub.1 and 410.sub.2 without providing for respective components for conversion of the provided analogue electronic signal into the digital signal.
[0179] According to an embodiment
[0180] That is, one, more or all modular recording sites of the probe may be configured for sampling a biosignal with at least a first and a second sensor element and for multiplexing outputs of the first and the second sensor element into the digital sensor signal. Alternatively or in addition to a time-multiplexing, a different or further concept of multiplexing may be implemented, e.g., a frequency-multiplex.
[0181]
[0182] The modular recording site 500 from
[0183] The quantizer 520 may comprise a reset 522, a clock 524 and a delay clock 526, wherein the delay clock 526 connects to the next side in chain (to the next modular recording site). The reset 522, for example, switches a transistor to reset the quantizer 520. Thus, for example, by the quantizer 520 a digitized sensor signal gets delayed in respect to the digital sensor signal of a neighboring modular recording site. The counter 530 is, for example, a ten-bit counter with a reset 532 and a clock 534. The counter 530 writes, for example, the digital sensor signal into a latch 540 (which has, for example, the function of the storage 460 from
[0184] In
[0185] The feedback current sinks (for example, the connection between the feedback output 528a and the feedback input 528b) are, for example, derived from the global bias line (control voltage 513). The feedback current, for example, determines the full scale (FS) of the ADC which can be configured to ±11.25 mV, ±22.5 mV or ±45 mV. Depending on the comparator output (for example, the feedback output 528a), the current may be injected either to the left or the right low-impedance cascade node of the OTA (integrator 510). Common mode ripple caused by the asymmetrical feedback may be reduced by connecting the 95 fF MIM integration capacitances (a first MIM (metal/insulator/metal) integration capacitance 519a and a second MIM integration capacitance 519b, wherein the first MIM integration capacitance 519a and the second MIM integration capacitance 519b each occupy, for example, less area than 20×10 μm.sup.2, 15×7 μm.sup.2 or 10×4 μm.sup.2, such as, for example, 7×3.5 μm.sup.2 and can have a capacitance in the scope of 20 fF to 200 fF, 50 fF to 150 fF or 80 fF to 100 fF, such as, for example, 95 fF) to V.sub.CMFB 590 and rejected by the differential comparator input. The noise of the feedback current source and the feedback switches, which may operate at digital-level input signals, is negligible compared to the major noise contributors. The constraints on the area and accuracy of the common-mode feedback is not stringent, first because noise is canceled by the differential nature of the circuit, and second, because no exact common mode is needed at the comparator input (for example, a first input 517a and a second input 518a). The trans-conductance of the differential pair may be determined by thermal noise considerations, e. g., to be in the scope of 1 μS to 20 μS, 2 μS to 10 μS or 3 μS to 5 μS, such as, for example, 4.2 μS. A measured maximal SNR may be less than 300 dB, 200 dB or 100 dB, such as, for example, 65.6 dB (FS=±45 mV) and a THD of, for example, less than 5%, 1% or 0.5%, such as, for example, 0.22% at V.sub.PP=10 mV (FS=±11.25 mV) is obtained for a tail current of 1.5 μA, according to this example.
[0186] In other words, the integrator 510 of each modular recording site of the plurality of modular recording sites is configured to receive the neuronal signal and to integrate the neuronal signal, so as to obtain an integrated neuronal signal. The quantizer 520 of each modular recording site of the plurality of modular recording sites comprises a latched comparator 570 and an output latch 580. The latched comparator 570 is configured to receive the integrated neuronal signal and to quantize the integrated neuronal signal. The output latch 580 is configured to drive switches for a feedback current 528 to the integrator 510, based on the comparator output latch 580. Depending on the comparator output latch 580, the current is injected either to the left or to the right low impedance cascade node of the OTA (the OTA together with a capacitance is an example for an integrator).
[0187]
[0188]
[0189]
[0190]
[0191]
[0192]
[0193] In the following, an example measurement is described shortly. The averaged data of all ADCs (analog-digital converter in each modular recording site of a neuronal probe according to an embodiment) is used to drive a proper body voltage for the application in saline solution and for cancellation of artifact signals. The measured signal is separated into low frequency local-field potentials (LFP) and high-frequency action potentials (AP) by digital post-processing.
[0194]
[0195] A pulsing (on/off with a frequency of 20 Hz or 1 kHz) broadband light source 810 (λ=400 nm to 1000 nm) may sends a beam 820 through a neutral density filter 830 onto an active needle probe 800 (the active needle probe 800 can, for example, have the same functionality as the neuronal probe 100, 200 and 300), wherein the active needle probe 800 is placed in a ringer solution 840. In the right diagram the first curve 850 corresponds to a high frequency (for example, 1 kHz) and a second curve 852 corresponds to a low frequency (for example, 20 Hz). In the diagram is also shown the full bandwidth noise limit 854. The shielding and layout concept (as described above) of the neuronal probe according to an embodiment with both input transistors of a differential pair (The differential transistor pair, describes a circuit technique in which two input transistors are used in order to create differential signaling in the signal path.) placed under the electrode suppresses illumination artifacts, which is a strong requirement for optogenetic stimulation of neuronal cells [9]. Sensitivity measurements against pulsing broadband light sources are shown in
[0196]
[0197] The neuronal probe 940 comprises an optional first electrode 942, wherein the first electrode (for example, a Pt-electrode) can be directly connected to V.sub.REF pad for electrode characterization. The neuronal probe 960 comprises Pads 962 and sensor elements 964, wherein the Pads 962 are, for example a contacting to connect the neuronal probe 960 to an external device. The neuronal probe 950 comprises Pads 962, sensor elements 964, a base 966 and an optional Si.sub.3N.sub.4—SiO.sub.2 passivation, wherein the Pads 962 are, for example a contacting to connect the neuronal probe 960 to an external device, the sensor elements 964 are, for example, Pt-electrodes (platinum-electrodes) and the Si.sub.3N.sub.4—SiO.sub.2 passivation forms an insulating portion 968 between each sensor element 964. The neuronal probe is separated after post-CMOS fabrication.
[0198] The micrographs show the fully implantable probe with, for example, a constant width of and thickness of 50 μm from tip to base. The length of the base is, for example, independent on the number of electrodes. The maximal number of recording sites may be solely limited by the data rate of the chain as the clock frequency equals to fs of the ADCs. Each ADC delivers, for example, 20 kS/s, limiting the length to, for example, 93 electrodes per chain. An example probe employs, for example, two data chains (each with, for example, 93 modular recording sites); however, an extension with multiple chains would only add marginal complexity to the digital part. Since no global analog neural signal routing may be present and due to the high modularity of the design, a longer probe or any application-specific modification of the probe geometry would deliver identical performance. Technology scaling would considerably reduce the power dissipation as well as the probe width, since half the probe area is dedicated to the digital circuitry.
[0199]
[0200]
[0201] With the neuronal probe as described herein a number of components of the signal processing unit 1130 can be reduced by omitting the signal amplifier 1132 and the analog/digital conversion 1134. For example the neuronal probe already includes those components in each modular recording side. Accordingly, it can be seen that the neuronal probe according to the present invention clearly outperforms conventional solutions.
[0202]
[0203] Whilst some embodiments are described in connection with an analog-to-digital conversion using an integrator and a comparator or quantizer, in a specific example in connection with the Gm-C based incremental delta-sigma analog-digital converter, i.e., a delta-sigma modulator of first order, embodiments described in the following further provide for an advantageous implementation of the analog-to-digital conversion by using a reconfigurable continuous-time incremental delta-sigma converter, i.e. an in-situ analog-to-digital converter 226 (see 226.sub.1 to 226.sub.n), that is configured for operating in a first operating mode 227a (see 227a.sub.1 to 227a.sub.n) for a coarse quantization and in a second operating mode 227b (see 227b.sub.1 to 227b.sub.n) for a fine quantization, i.e., for quantizing a remaining error 225 (see 225.sub.1 to 225.sub.n) of the first operating mode 227a, e.g., see
[0204] The first operating mode 227a and the second operating mode 227b may differ from each other, e.g., in view of the amplification factor implemented in a feedback of the delta-sigma converter, in view of a sampling rate used and/or in view of a signal shape or signal type used for a feedback digital-to-analog converter 226.
[0205] A sensor array described in the following might comprise features and or functionalities as described with regard to one of the embodiments described above. Those embodiments may be combined without limitation with the embodiments described above, e.g., to substitute the prior ADC elements or to modify the analog-to-digital conversion. However, the advantage of reusing elements of a delta-sigma modulator in different operating modes 227 (see 227a and 227b) to quantize a single analog signal 132 (see 132.sub.1 to 132.sub.n), the results of the different operating modes 227 to be combined with each other to obtain a combined result for the analog signal 132 is not limited to neural probes or even probes but may be used in any other application using an analog-to-digital conversion, e.g., sensor applications, communication applications or the like.
[0206]
[0207] Each modular recording site 2200 of the plurality of modular recording sites 2200 comprises, a CMOS substrate 2210 (see 2210.sub.1 to 2210.sub.n), at least one sensor element 2220 (see 2220.sub.1 to 2220.sub.n), an in-situ analog-to-digital converter 226, and a communication interface 2230 (see 2230.sub.1 to 2230.sub.n).
[0208] The electrical components of a modular recording site 2200, like the in-situ analog-to-digital converter 226 and the communication interface 2230, might be integrated in the CMOS substrate 2210 and might be positioned under a surface of the at least one sensor element 2220. A housing into which the surface of the at least one sensor element 2220 might be integrated can shield the electronic components from external influences.
[0209] The at least one sensor element 2220 is configured for receiving, e.g. detecting, an analog signal 132. For example, in the surroundings of the sensor array 2000 a biosignal 130 might occur, which can be detected by at least one of the pluralities of modular recording sites 2200. A sensor element 2220 of a modular recording site 2200 might detect the biosignal 130 as an analog signal 132, wherein the analog signal 132 represents the biosignal 130 at the position of the respective modular recording site 2200, e.g. with the same characteristics but with a damped amplitude due to a propagation loss. The respective at least one sensor element 2220 might be configured to forward the analog signal 132 directly to the in-situ analog-to-digital converter 226 or to a pre-processing means configured for providing a processed analog signal to the in-situ analog-to-digital converter 226.
[0210] The respective in-situ analog-to-digital converter 226 is configured for converting the analog signal 132 into a digital sensor signal 134 (see 134.sub.1 to 134.sub.n). Each in-situ analog-to-digital converter 226 is configured for operating in the first operating mode 227a for performing a first quantization of the respective analog signal 132 using a first quantization setting and to obtain a residual error 225 from the first quantization; and for operating in a second operating mode 227b for performing a second quantization of the residual error 225 using a second, different quantization setting for a same element of the respective in-situ analog-to-digital converter 226. The in-situ analog-to-digital converter 226 might be configured to provide the digital sensor signal 134 to further processing means or directly to the communication interface 2230.
[0211] The communication interface 2230 is configured to provide the digital sensor signal 134 or a further processed digital sensor signal to the base 2100. The communication interfaces 2230 of the plurality of modular recording sites 2200 are connected serially with respect to each other and to the base 2100.
[0212] The base 2100 is configured to provide a probe signal 112. The probe signal 112 might correspond to the digital sensor signal 134 provided by the communication interface 2230. Alternatively, the base 2100 might be configured to determine the probe signal 112 based on the digital sensor signal 134 provided by the communication interface 2230.
[0213] The sensor array 2000 might comprise features and/or functionalities which are described in the following with regard to
[0214] Embodiments provide for a small, e.g., having an area of 0.00378 mm.sup.2, scalable neural recording front-end for fully-immersible neural probes based on a two-step incremental delta-sigma converter 226 with extended counting and hardware reuse.
[0215] Embodiments relate, thus, to a delta-sigma modulator which is incorporated into a neuronal probe according to some embodiments. For example, such a delta-sigma modulator may be included in a modular recording site 2200 or at a different location, e.g., the base 2100.
[0216] According to further embodiments, the analog-to-digital converter 226 may be used without a neuronal probe.
[0217] A neuronal probe may be implemented such at that one, or more than one or every analog-to-digital converter 226, e.g., at each modular recording site 2200, the in-situ analog-to-digital converter 226 is configured for operating in a first operating mode 227a for performing a first quantization of the biosignal, i.e. the analog signal 132, using a first quantization setting and to obtain a residual error 225 from the first quantization; and for operating in a second operating mode 227b for performing a second quantization of the residual error 225 using a second, different quantization setting for a same element of the analog-to-digital converter 226. Thus, for the analog-to-digital conversion, an element, like an amplification element and/or a sampling element, can be used in both operating modes 227 by applying different settings dependent on the mode. This results in a small area needed for all components of the analog-to-digital converter 226, since an element, two or more elements or all elements used in the first operating mode 227a can be reused in the second operating mode 227b.
[0218] According to an embodiment, the quantization setting corresponds to an amplification or a gain applied at the analog-to-digital conversion. An example for a first amplification or gain may be 1, less than one or more than 1. An example for an amplification, gain respectively of the second quantization setting may be a different value, e.g., a lower value such as 1/16, ⅛ or ¼ with regard to 1 or the amplification value of the first quantization setting.
[0219] According to an embodiment, the quantization setting corresponds to a sampling rate applied at the analog-to-digital conversion. An example for a first sampling rate of the first quantization setting may comprise a normalized value of 1 and/or an absolute value of 1 MHz or less, at least 1.5 MHz or at least 2 MHz such as 2.72 MHz or even more. An example second sampling rate of the second quantization setting may comprise a normalized value of at least 5, at least 8 or more with regard to the first sampling rate, e.g., 21.76 MHz which may be 8 times 2.72 MHz. For example, a second gain being ⅛ of the first gain may be selected in accordance with 8 times a sampling rate of the second setting with regard to the first setting, rendering 8 as a relationship value. Alternatively, a different relationship value may be selected or no relationship value may be selected.
[0220] According to an embodiment, the quantization setting corresponds to a signal shape applied for sampling at the analog-to-digital conversion. An example first signal of the first quantization setting may comprise a non-return-to-zero pulse shape feedback signal, whilst the second signal of the second quantization setting may comprise a return-to-zero feedback signal, e.g., having a duty cycle of 12.5% and, thus, corresponding to the relationship value.
[0221] According to an embodiment, the first quantization setting may comprise an amplification of 1, a sampling rate of 2.72 MHz; a signal shape as a non-return-to-zero pulse shape feedback signal; and the second quantization setting may comprise an amplification/gain of ⅛, a sampling rate of 21.76 MHz; a signal shape as a return-to-zero feedback signal with 12.5% duty cycle to control the element “digital-to-analog converter”, DAC, of the analog-to-digital converter. Different values may be selected for one or more parameters.
[0222] According to an embodiment, see e.g.
[0223] According to an embodiment the first operating mode 227a and the second operating mode 227b differ from each other, in view of an amplification 2245a/2245b applied in a feedback loop of a delta-sigma modulator 2250 of the analog-to-digital converter 226; a sampling rate 2248 of the delta-sigma modulator 2250; and/or of a pulse shape of the feedback DAC 2246a/2246b in the delta-sigma-modulator 2250.
[0224] For example, as pulse shapes of the feedback DAC 2246a/2246b non-return-to-zero and return-to-zero may be used in embodiments.
[0225]
[0226] Assuming an amplification 2245a of 1 in the feedback path of the two-step incremental delta-sigma converter, i.e. the in-situ analog-to-digital converter 226, with M.sub.2=8 as well as an sampling frequency 2248 of 2.72 MHz and a non-return-to-zero feedback signal in the first operating mode 227a, the amplification 2245b is reduced by the factor M.sub.2, i.e., the amplification 2245b is ⅛, for fine quantization in the second operating mode 227b or the sampling frequency 2248 is increased by the factor of M.sub.2, i.e., the sampling frequency 2248 is 21.76 MHz, or a return-to-zero signal with a duty cycle of 100%/M.sub.2 is used, i.e., the duty cycle is 12.5%, or a combination of the mentioned methods, e.g., the amplification 2245b in the second phase, i.e. in the second operating mode 227b, is ¼ and the sampling frequency 2248 is 5.44 MHz or the amplification 2245b is ¼ and the duty cycle of the return-to-zero feedback signal is 50%.
[0227] According to an embodiment, in the second operating mode 227b an integrator 2260 and/or a quantizer 2270 and/or a feedback DAC 2246 of a delta-sigma modulator 2250 of an analog-to-digital converter 226 is reused with respect to the first operating mode 227a. In case of reusing the feedback DAC 2246a, as shown in
[0228] In other words:
[0229] Embodiments provide for an analog front-end for recording neural signals based on a multi-step continuous-time incremental delta-sigma analog-to-digital converter.
[0230] The architecture of the analog front-end is based on a multi-step continuous-time IΔΣ ADC, i.e. the in-situ analog-to-digital converter 226, that combines the advantages of first-order and higher-order ΔΣ systems. This means that, compared to a conventional first-order system, the oversampling ratio can be reduced by using a multi-step method with coarse and fine quantization. At the same time, the area can be maintained since it is sufficient for the system to use one integrator 2260 that can be reused for the different stages, i.e. for the first operating mode 227a and for the second operating mode 227b, of the quantization process. The system architecture of the continuous-time ADC is shown for the different phases, i.e. for the first operating mode 227a and the second operating mode 227b, of the quantization process in
[0231] To ensure that only the remaining quantization error from phase 1 227a remains in the integrator at the beginning of phase 2 227b, the input signal u.sub.in(t) 132 is disconnected 2244 in the last clock cycle of phase 1 227a (see
[0232]
[0233] During this phase, i.e. the second operating mode 227b, the stored value in the integrator 2260 y(t) is integrated towards the center value of the ADC. In doing so, the number of 1s and 0s occurring at the output of the comparator 2270 is counted. The result is added to the result of the coarse quantization. Furthermore, the remainder of the fine quantization is evaluated in the last clock cycle of phase 2 227b, gaining one additional bit of resolution. To this end, the sign of the residual value y(t) is checked and the result is attached to the previous one. Subsequently, the modulator 2250 is reset, and a new analog-to-digital conversion can be started. The resolution of the system, i.e. the in-situ analog-to-digital converter 226, is given by the number of clock cycles in the respective phases 227a and 227b with ENOB=log.sub.2(M.sub.1)+log.sub.2(M.sub.2)+1.
[0234] The two-step architecture in
[0235] The front-end architecture in
[0236]
[0237] The core of the transistor-level implementation is the gm-C integrator 2260 shown in
[0238] With the multi-step quantization method, the oversampling ratio can be reduced by a factor of (2M.sub.1M.sub.2)/(M.sub.1+M.sub.2) compared to a conventional first order IΔΣ system with the same resolution, leading to a strongly reduced power consumption on a similar silicon area due to the hardware reuse in the second operating mode 227b. This results in a greatly reduced power consumption, and since the integrator 2260 and the comparator 2270 are used for all phases 227a and 227b, the silicon area does not increase. Due to the reduced power, time-multiplexing methods can be employed without violating temperature requirements. With such a method, the area per channel could be further reduced. In contrast to discrete-time systems, the continuous-time implementation based on a gm-C integrator 2260 makes it possible to realize the system, e.g. the sensor array 2000, with a small integration capacitor on a smallest possible area and with low noise. Due to the intrinsic low-pass behavior of the ADC, an additional anti-aliasing filter becomes obsolete. Furthermore, cross-coupling the input transistors saves area by avoiding the need to add additional input transistors or a multiplexer in front of the input to disconnect 2244 the input signal 132. Additionally, it avoids the need to switch at the high-impedance input node Vin.
[0239] The herein proposed front-end architecture has the advantage that the area, the noise, and the power consumption could be strongly reduced compared to other concepts.
[0240] In again other words, a front-end having a possible size of 0.00378 mm.sup.2 and, thus a scalable neural recording front-end for fully immersible neural probes based on a two-step incremental delta-sigma converter 226 with extended counting and hardware reuse is described.
[0241] The extensive integration of electronics into tissue-penetrating probes improves the signal quality and reduces parasitic effects for high-density recording of in vivo neural activity. In contrast to passive neural probes or devices implementing only part of the signal chain in the probe shank [5], [6], [11], fully immersible subcortical probes allow the recording of neural signals in deep brain regions [12]. This is achieved by directly digitizing brain activity in situ, thus avoiding a large base and allowing the probe to have a base and shank of equal width. However, this comes with a lower spatial resolution and an increased power density in the probe shank. To advance the concept of fully immersible probes, neural recording front-end architectures are required which reduce not only the area, but also the power per channel, thus avoiding tissue overheating due to increased power density. This paper presents a modular neural recording front-end, i.e. the sensor array 2000, which achieves these goals while also enhancing the noise and linearity performance compared to the state of the art.
[0242]
[0243] The operation of the two-step converter front-end shown in
[0244]
[0245]
y(M.sub.1)=b.sub.1∫.sub.0.sup.M.sup.
[0246]
[0247]
[0248]
[0249] The differential output current for the first operating mode 227a can be determined according to I.sub.out=Gm.sub.in.Math.(V.sub.electrode−V.sub.BODY) and for the second operating mode 227b according to I.sub.out≈0.
[0250] The IDAC 2246 for offset compensation in
[0251]
[0252]
[0253]
[0254] The functionality of the proposed system is validated in saline solution on an 8-channel prototype with offset compensation by the experimental setup shown in
[0255] The presented modular neural recording front-ends are implemented in a 0.18 μm CMOS technology on an area of 0.00378 mm.sup.2 without and 0.00462 mm.sup.2 with offset compensation. This results in E-A Ch. FoMs [11] of 4.04 fJ/C.Math.s.Math.mm.sup.2 and 5.97 fJ/C.Math.s.Math.mm.sup.2 for SNDRs of 58.04 dB and 57 dB. Both versions are compared to the state-of-the-art neural probes in
[0256]
[0257]
Conclusions:
[0258] Neural Recording Front End based on a Two Step IΔΣ ADC [0259] Reduced OSR and thus power consumption at the same resolution of a first-order IΔΣ ADC [0260] Minimal silicon area due to hardware reuse [0261] Fully Immersible Neural Probe Prototypes [0262] ±60 mV offset compensation/4-bit configurable [0263] Lowest power consumption for fully integrated neural probes [0264] Best noise performance for fully integrated neural probes [0265] Highly scalable with process technology and supply since up to 67% of the circuit is covered by digital cells
[0266]
[0267] The base 2100 is configured to provide a probe signal 112 based on the digital sensor signals 134 provided by the plurality of modular recording sites 2200. The base 2100 might receive from each modular recording site 2200 the respective digital sensor signals 134 and can be configured to provides same as the probe signal 112. Alternatively, the digital sensor signals 134 are further processed to obtain the probe signal 112.
[0268] Further, the sensor array 2000 comprises a data compression unit 4000, e.g. a data compressor, configured for reducing a data rate of the digital sensor signal 134.sub.1 obtained by a modular recording site 2200.sub.1 of the plurality of modular recording sites 2200. The in-situ analog-to-digital converter 226 might be configured to directly provide the digital sensor signal 134.sub.1 to the data compression unit 4000. Alternatively, the modular recording site 2200 and/or the base 2100 might comprise additional processing means for processing the digital sensor signal 134.sub.1 before providing it to the data compression unit 4000.
[0269] As shown in
[0270] According to an embodiment, the data compression unit 4000 is configured for determining a difference 4100a between a first digital sensor signal 134a.sub.1 obtained by the modular recording site 2200.sub.1 during a first instance of time t.sub.1 and a second digital sensor signal 134b.sub.1 obtained by the modular recording site 2200.sub.1 during a second, later instance of time t.sub.2, see
[0271] According to an alternative embodiment, also shown in
[0272] According to an embodiment, the base 2100 is configured to provide the probe signal 112 based on the delta-signal, e.g., the difference 4100a, or the delta-value, e.g., the difference 4100b. Advantageously, the base 2100 is configured to provide the delta-signal, e.g., the difference 4100a, or the delta-value, e.g., the difference 4100b, as the probe signal 112.
[0273] According to an embodiment, the data compression unit 4000 integrated in the base 2100 can be configured to receive from each modular recording site 2200 of the plurality of modular recording sites 2200 the respective digital sensor signal 134 and reduce a respective data rate of the respective digital sensor signal 134. For each digital sensor signal 134 provided by the plurality of modular recording sites 2200, the data compression unit 4000 can be configured to obtain the difference 4100a or 4100b as described above. The difference 4100a, for example, is determined between digital sensor signals 134 obtained by and provided from the same modular recording site 2200 of the plurality of modular recording sites 2200, e.g., a difference 4100a between two consecutive digital sensor signals 134 received from the same modular recording site 2200.
[0274] According to another embodiment, a first modular recording site 2200.sub.1 of the plurality of modular recording sites 2200 comprises the data compression unit 4000 and a second modular recording site 22002 of the plurality of modular recording sites 2200 comprises a further data compression unit. The further data compression unit can comprise features and/or functionalities as described above with regard to the data compression unit 4000.
[0275] According to an embodiment, each modular recording site 2200 of the plurality of modular recording sites 2200 can comprise a data compression unit with the same features and/or functionalities as described above with regard to the data compression unit 4000. Thus, each modular recording site 2200 can provide the respective digital sensor signal 134 with a reduced data rate to the base 2100.
[0276] According to an embodiment, each modular recording site 2200 of the plurality of modular recording sites 2200 can further comprise a communication interface (e.g., 228, 328, 440, 460, 550 and/or 2230) as described with regard to one of the aforementioned sensor arrays. The respective communication interface might be configured to provide the respective digital sensor signal 134 or the digital sensor signal 134 with a reduced data rate to the base 2100, dependent on whether the respective modular recording site comprises a data compression unit 4000 or not.
[0277] This sensor array 2000 enables a reduction of the data rate in fully-integrated CMOS neuronal probes with local digitization based on digital delta encoding. An analysis of neuronal signals, i.e. analog signals 132, recorded during in-vivo experiments revealed that the data compression, e.g. delta encoding, performed by the data compression unit 4000 can reduce the data rate by 36% with respect to a full scale of ±11.25 mV and a resolution of 11 bit, e.g. of the in-situ analog-to-digital converter 226.sub.1. The delta encoding can be determined off-chip or directly integrated into a decimation filter of an in-situ analog-to-digital converter 226.sub.1, e.g., an incremental delta-sigma ADC, located under each sensor element 2220, e.g. under the electrodes, on the shank of the probe, i.e. the sensor array 2000. An area estimate of the filter by synthesis in a 180 nm technology resulted in a 34% reduced area when using delta encoding.
[0278] Optionally, each of one or more modular recording sites 2200 of the plurality of modular recording sites 2200 further comprises a reduction element 4200 as will be described in more detail with regard to
[0279]
[0280] According to an embodiment, the data compression unit 4000 of the sensor array 2000 or a plurality of data compression units comprising the data compression unit 4000 is configured for determining, for each digital sensor signal 134 provided by the plurality of modular recording sites 2200, a respective difference 4100a or 4100b, wherein respective raw data, e.g. a reconstructed value, a reconstructed signal, a decoded value or a decoded signal, representing the respective originally received analog sensor signal 132 is determinable based on the respective difference 4100a or 4100b.
[0281] To control the potential of a reference electrode in the DC-coupled neuronal probe 2000, the uncompressed raw data, for example, are needed. Since delta coding is lossy, an initial reference sweep is presented to enable regulation. This is validated with measurements in a phosphate buffered saline.
[0282] According to an embodiment, in an initial step the reference control unit 3100 is configured to change the reference voltage of the reference electrode 3150 until, for each modular recording site 2200 of the plurality of modular recording sites 2200, the respective difference 4100a or 4100b equals zero. Then, in a further step, the reference control unit 3100 is configured to further change the reference voltage of the reference electrode 3150 with a configurable increment and obtain, for each reference voltage, an average of a plurality of the reconstructed values, and perform the further change of the reference voltage until the average of the plurality of the reconstructed values is equal to a predetermined value. It is possible that at an initialization of the initial step the delta values associated with the plurality of modular recording sites 220 are already all equal to zero. In this case, the reference control unit 3100 is configured to continue with the further step.
[0283] More optional details with regard to the measurement system 3000 and the reference sweep are described with regard to
[0284]
[0285] Each modular recording site 2200 of the plurality of modular recording sites further comprises a reduction element 4200 configured to reduce a word size of the respective digital sensor signal 134 provided by the respective in-situ analog-to-digital converter 226 of the respective modular recording site 2200. For example, for each modular recording site 2200, the respective reduction element 4200 can be configured to provide the respective digital sensor signals with a word size smaller than a resolution of the respective in-situ analog-to-digital converter 226. The reduced word size corresponds to a first number 4210a of bits being smaller than a second number 4210b of bits representing the resolution of the in-situ analog-to-digital converter 226, e.g. a 6 bit reduced word size at an 11 bit resolution.
[0286] According to an embodiment, for each modular recording site 2200, the respective reduction element 4200 is configured to reduce the word size of the respective digital sensor signal 134 by omitting a predefined number of high-order bits, e.g., omitting the certain number of high-order bits so that the respective digital sensor signals 134 are represented by at least 6 lower-order bits.
[0287] The herein proposed sensor arrays, especially the one described with regard to
[0288] With neuronal signals, a differentiation is conventionally made between low-frequency local field potentials (LFP, 0.5 Hz-1 kHz) and high-frequency action potentials (AP, 300 Hz-10 kHz), see [12], [16-18]. Although the value ranges of the read-out electronics of the probe are specified with ±11.25 mV, ±22.5 mV and ±45 mV in [12], they are not designed for pure neuronal signals, but they are designed to be able to cover variations of the electrochemical potential difference between the individual electrodes. However, the maximum amplitudes of the neuronal signals measurable by means of invasive methods, e.g. with a neuronal probe, are specified with 1 mV [20]. Thus, the full dynamic range of the read-out system is not required for the signals of interest, promising possible compression of the digitized neuronal signals through consideration of the delta values.
[0289] In order to be able to make a statement as to whether delta encoding is a suitable compression method for neuronal read-out systems with in-situ ΔΣ ADC front-end, and as to which word width, i.e. word size, in bits is sufficient for a delta data transfer, data sets from in-vivo experiments are statistically analyzed with respect to their delta values first. The statistical parameters (standard deviation a and maximum delta values Δ.sub.max), the value range (FS) and the least-significant bit (LSB) of the evaluated raw data with a sampling rate of 20 kHz are given in Table 1. The data sets nc-m1, nc-m2, and nc-m3 were recorded with fully immersible CMOS probes [12] in the motor cortex of non-human primates, and the data set HC2 (ec013.527) [21] was recorded in the hippocampus of rats.
TABLE-US-00001 TABLE 1 Statistical analysis of raw data from available in-vivo experiments recorded at a sampling rate of 20 kHz. data set FS (mV) LSB (μV) σ (μV) Δ.sub.max (μV) nc-m1 ±11.25 11 20.3 275 nc-m2 ±22.5 22 22.9 176 nc-m3 ±45 44 27.5 176 hc2 20 0.3 9.32 343
[0290] However, the standard deviation is not meaningful in this case, since, even when considering a range of 12σ, not all of the maximum delta values are reached. The latter are associated with the APs, which becomes apparent through digital post-processing of the raw data, i.e. band-pass filtering in the corresponding frequency range. Due to the refractory period of a neuron, APs occur only sporadically with a frequency limit of approximately 500 Hz; however, due to their characteristic shape they have larger rates of change. On the other hand, LFPs only change slowly, which is why the standard deviation is low and the maximum delta value is large in comparison. With delta encoding, it has to be ensured that the maximum signal change can be detected. According to the data evaluation in Table 1, this is possible for a maximum delta value of 343 μV with a 6-bit data word plus a sign bit, i.e. with a word size corresponding to a number of bits being 6 plus the sign bit. In this case, reference is made to a probe, i.e. a sensor array 2000, with a value range of ±11.25 mV and a resolution of 11 bits [12]. Accordingly, signal change rates of approximately 700 μV/sample can be detected, and the data rate can be reduced by 36%.
[0291]
[0292] However, delta encoding is a lossy compression method, only containing information about time-varying signals. For AC-coupled read-out systems [16, 17], this loss of information would not be relevant, however, this information is required for DC-coupled systems with such a reference closed-loop control [12] and [18]. Information with respect to a DC portion or an initial offset is lost, said information being essential to correctly closed-loop control the compensation voltage. In order to ensure the functionality of the reference closed-loop control in combination with delta encoding, an initial reference sweep 3500 is performed according to the flow diagram in
[0293] In other words, at the reference sweep 3500, the reference control unit 3100 is configured to control the reference voltage of the reference electrode 3150 by [0294] applying in an initial step 3505 a predetermined reference voltage to the reference electrode 3150, [0295] obtaining from the data compression unit 4000, for each modular recording site 2200 of the plurality of recording sites 2200, the respective difference 4100, so that a plurality of differences 4100 is obtained, [0296] checking 3510 whether all differences 4100 of the plurality of differences 4100 are equal to zero, and [0297] if all differences of the plurality of differences are equal to zero, the reference control unit 3100 is configured to [0298] change 3540 the reference voltage in a first direction with a configurable increment until at least one difference 4100 unequal to zero is obtained from the data compression unit 4000, and [0299] then continue changing 3530 the reference voltage in the first direction and obtain an average of signals reconstructed based on the plurality of differences 4100 determined by the data compression unit 4000 until the average is equal to a predetermined value, and optionally [0300] then automatically perform the reference control (e.g., P, I, D, PI, PD, or PID control), i.e., direction and increment are no longer predetermined; [0301] if at least one of the plurality of differences 4100 is unequal to zero, the reference control unit 3100 is configured to [0302] change 3520 the reference voltage in a second direction with a configurable increment until for each modular recording site 2200 of the plurality of modular recording sites 2200, a difference 4100 equal to zero is obtained from the data compression unit 4000, and [0303] then change 3530 the reference voltage in a direction opposite to the second direction and obtain an average of signals reconstructed based on the plurality of differences 4100 determined by the data compression unit 4000 until the average is equal to the predetermined value, and optionally [0304] then automatically perform the reference control (e.g., P, I, D, PI, PD, or PID control), i.e., direction and increment are no longer predetermined.
[0305] As is illustrated in
[0306] To validate the concept of the delta transfer with reference closed-loop control and the initial reference sweep, the concept described in connection with
[0307] In case of the probe 2000 used, since a decimation filter 2290 is integrated on the shank in each of the 144 channels 2200, the 11 bits of raw data 134 were initially transferred from the probe 2000 to the FPGA, and the delta values 4100 for controlling the reference voltage were determined there. The encoded and decoded data as well as the voltage at the reference electrode 3150 during the initial reference sweep 3500 are illustrated in the measurement results in
[0308] Although some aspects have been described in the context of an apparatus, it is clear that these aspects also represent a description of the corresponding method, where a block or device corresponds to a method step or a feature of a method step. Analogously, aspects described in the context of a method step also represent a description of a corresponding block or item or feature of a corresponding apparatus. Some or all of the method steps may be executed by (or using) a hardware apparatus like, for example, a microprocessor, a programmable computer or an electronic circuit. In some embodiments, one or more of the most important method steps may be executed by such an apparatus.
[0309] Depending on certain implementation requirements, embodiments of the invention can be implemented in hardware or in software. The implementation can be performed using a digital storage medium, for example, a floppy disk, a DVD, a Blu-Ray, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, having electronically readable control signals stored thereon, which cooperate (or are capable of cooperating) with a programmable computer system such that the respective method is performed. Therefore, the digital storage medium may be computer readable.
[0310] Some embodiments according to the invention comprise a data carrier having electronically readable control signals, which are capable of cooperating with a programmable computer system, such that one of the methods described herein is performed.
[0311] Generally, embodiments of the present invention can be implemented as a computer program product with a program code, the program code being operative for performing one of the methods when the computer program product runs on a computer. The program code may, for example, be stored on a machine-readable carrier.
[0312] Other embodiments comprise the computer program for performing one of the methods described herein, stored on a machine-readable carrier.
[0313] In other words, an embodiment of the inventive method is, therefore, a computer program having a program code for performing one of the methods described herein, when the computer program runs on a computer.
[0314] A further embodiment of the inventive methods is, therefore, a data carrier (or a digital storage medium, or a computer-readable medium) comprising, recorded thereon, the computer program for performing one of the methods described herein. The data carrier, the digital storage medium or the recorded medium are typically tangible and/or non-transitionary.
[0315] A further embodiment of the inventive method is, therefore, a data stream or a sequence of signals representing the computer program for performing one of the methods described herein. The data stream or the sequence of signals may, for example, be configured to be transferred via a data communication connection, for example, via the Internet.
[0316] A further embodiment comprises a processing means, for example, a computer, or a programmable logic device, configured to or adapted to perform one of the methods described herein.
[0317] A further embodiment comprises a computer having installed thereon the computer program for performing one of the methods described herein.
[0318] A further embodiment according to the invention comprises an apparatus or a system configured to transfer (for example, electronically or optically) a computer program for performing one of the methods described herein to a receiver. The receiver may, for example, be a computer, a mobile device, a memory device or the like. The apparatus or system may, for example, comprise a file server for transferring the computer program to the receiver.
[0319] In some embodiments, a programmable logic device (for example, a field programmable gate array) may be used to perform some or all of the functionalities of the methods described herein. In some embodiments, a field programmable gate array may cooperate with a microprocessor in order to perform one of the methods described herein. Generally, the methods are performed by any hardware apparatus.
[0320] The apparatus described herein may be implemented using a hardware apparatus, or using a computer, or using a combination of a hardware apparatus and a computer.
[0321] The apparatus described herein, or any components of the apparatus described herein, may be implemented at least partially in hardware and/or in software.
[0322] The methods described herein may be performed using a hardware apparatus, or using a computer, or using a combination of a hardware apparatus and a computer.
[0323] The methods described herein, or any components of the apparatus described herein, may be performed at least partially by hardware and/or by software.
[0324] While this invention has been described in terms of several embodiments, there are alterations, permutations, and equivalents which fall within the scope of this invention. It should also be noted that there are many alternative ways of implementing the methods and compositions of the present invention. It is therefore intended that the following appended claims be interpreted as including all such alterations, permutations and equivalents as fall within the true spirit and scope of the present invention.
REFERENCES
[0325] [1] G. Buzsaki et al., “Tools for Probing Local Circuits: High-Density Silicon Probes Combined with Optogenetics,” Neuron, vol. 86, pp. 92-105, 2015. [0326] [2] K. Seidl, M. Schwaerzle, I. Ulbert, H. P. Neves, O. Paul and P. Ruther, “CMOS-Based High-Density Silicon Microprobe Arrays for Electronic Depth Control in Intracortical Neural Recording—Characterization and Application,” in Journal of Microelectromechanical Systems, vol. 21, no. 6, pp. 1426-1435, December 2012. [0327] [3] A. S. Herbawi et al., “High-density CMOS neural probe implementing a hierarchical addressing scheme for 1600 recording sites and 32 output channels,” Transducers, pp. 20-23, 2017. [0328] [4] V. Viswam et al., “High-density mapping of brain slices using a large multi-functional high-density CMOS microelectrode array system,” Transducers, pp. 135-138. 2017, doi: 10.1109/TRANSDUCERS.2017.7994006 [0329] [5] C. M. Lopez et al., “22.7 A 966-electrode neural probe with 384 configurable channels in 0.13 μm SOI CMOS,” ISSCC Dig. Tech. Papers, pp. 392-393, 2016. [0330] [6] B. C. Raducanu et al, “Time multiplexed active neural probe with 678 parallel recording sites”, IEEE ESSDERC, 2016, pp. 385-388. [0331] [7] F Heer, W Franks, A Blau, S Taschini, C Ziegler, A Hierlemann, H Baltes, “CMOS microelectrode array for the monitoring of electrogenic cells”, Biosensors and Bioelectronics, vol. 20, pp. 358-366, 2004, ISSN 0956-5663, doi: 10.1016/j.bios.2004.02.006. [0332] [8] J. Scholvin et al., “Close-packed silicon microelectrodes for scalable spatially oversampled neural recording”, IEEE Trans. Biomed. Eng., vol. 63, pp. 120-130, 2016. [0333] [9] T. D. Y. Kozai and A. L. Vazquez, “Photoelectric Artefact from Optogenetics and Imaging on Microelectrodes and Bioelectronics: New Challenges and Opportunities,” J. Mater. Chem. B, pp. 4965-4978, 2015. [0334] [10] S. Tao and A. Rusu, “A Power-Efficient Time-continuous Incremental Sigma-Delta ADC for Neural Recording Systems,” in IEEE Transactions on Circuits and Systems I: Regular Papers, vol. 62, no. 6, pp. 1489-1498, June 2015, doi: 10.1109/TCSI.2015.2418892. [0335] [11] S. Park et al., “Modular 128-Channel—Analog Front-End Architecture Using Spectrum Equalization Scheme for 1024-Channel 3-D Neural Recording Microsystems”, IEEE JSSC, vol. 53, no. 2, pp. 501-514, February 2018. [0336] [12] D. De Dorigo et al., “Fully Immersible Subcortical Neural Probes With Modular Architecture and a Delta-Sigma ADC Integrated Under Each Electrode for Parallel Readout of 144 Recording Sites,” IEEE JSSC, vol. 53, no. 11, pp. 3111-3125, 2018. [0337] [13] Patent: WO2019154989—SENSOR ARRAYS, METHOD FOR OPERATING A SENSOR ARRAY AND A COMPUTER PROGRAM FOR PERFORMING A METHOD FOR OPERATING A SENSOR ARRAY [0338] [14] Patent: U.S. Pat. No. 5,410,310—METHOD AND APPARATUS FOR EXTENDING THE RESOLUTION OF A SIGMA-DELTA TYPE ANALOG TO DIGITAL CONVERTER [0339] [15] Y. Zhang et al., “A 16 b Multi-Step Incremental Analog-to-Digital Converter With Single-Opamp Multi-Slope Extended Counting”, IEEE JSSC, vol. 52, no. 4, pp. 1066-1076, April 2017. [0340] [16] C. M. Lopez et al., A Neural Probe With Up to 966 Electrodes and Up to 384 Configurable Channels in 0.13 μm SOI CMOS, IEEE Trans. Biomed. Circuits Syst., vol. 11, no. 3, pp. 510-522, June 2017 [0341] [17] B. C. Raducanu et al., Time Multiplexed Active Neural Probe with 1356 Parallel Recording Sites. Sensors (Basel, Switzerland), 2017 Oct. 19; 17(10):2388. [0342] [18] D. Wendler et al., 28.7 A 0.00378 mm2 Scalable Neural Recording Front-End for Fully Immersible Neural Probes Based on a Two-Step Incremental Delta-Sigma Converter with Extended Counting and Hardware Reuse, 2021 IEEE International Solid-State Circuits Conference (ISSCC), 2021, pp. 398 [0343] [19] S. Kim et al., Thermal impact of an active 3-D microelectrode array implanted in the brain, IEEE Trans. Neural Syst. Rehabil. Eng., vol. 15, no. 4, pp. 493-501, 2007 [0344] [20] Rikky Muller, Brain-computer interfaces: Fundamentals to future technologies, 2021 IEEE International Solid-State Circuits Conference (ISSCC) Tutorial [0345] [21] Collaborative Research in Computational Neuroscience—Data sharing, web: https://crcns.org/(besucht am 03/02/2021)