IMPLANTABLE MICRO DEVICE WITH HIGH DATA RATE BACK SCATTERING

Abstract

An implantable micro device, or dust, has a piezoelectric transducer connected to a power management circuit, which provides electric power output for powering components of the device based on an ultrasonic power signal from an external ultrasonic signal source. A sensor measures a physical parameter, e.g. a neural activity signal, and generates an electric signal, digitized by a time-encoding analog-to-digital converter, e.g. a delta-sigma modulator, to generate a one-bit data stream representing the sensed parameter. A load modulation circuit with one or more electric switches connected to the transducer modulates transducer electric load according to the one-bit data stream, thus causing a backscattered signal from the piezoelectric transducer to be modulated by the sensed physical parameter. Preferably, the piezoelectric transducer's electric load is harshly modulated by connecting it to either an optimum load for minimum reflection, or short-circuiting for maximum reflection according to each bit of the digitized data stream.

Claims

1-16. (canceled)

17. A micro device, arranged for implantation into biological tissue, the micro device comprising: a piezoelectric transducer; a power management circuit connected to the piezoelectric transducer, and being arranged to generate an electric power output for powering components of the micro device in response to an ultrasonic power signal received by the piezoelectric transducer from an external source; a sensor arranged to measure a physical parameter or a neural activity, and to generate an electric signal accordingly; an electric circuit arranged to receive the electric signal from the sensor, and to digitize the electric signal by means of time-encoding analog-to-digital converter, to generate a one-bit data stream representing the electric signal from the sensor; and a load modulation circuit comprising at least one electric switch connected to terminals of the piezoelectric transducer, so as to allow modulation of electric load of the piezoelectric transducer in response to the one-bit data stream, so that a backscattered signal from the piezoelectric transducer is modulated by the one-bit data stream.

18. The micro device according to claim 17, wherein the load modulation circuit is arranged to control the at least one electric switch according to the one-bit data stream at a modulation frequency.

19. The micro device according to claim 17, arranged to store a time sequence of the generated one-bit data stream in a memory, and applying the stored time sequence of the one-bit data stream to the load modulation circuit at an increased data rate to provide a time compression of data represented in the backscattered signal.

20. The micro device according to claim 19, wherein the increased data rate is at least a factor of 5, compared to a data rate of the one-bit data stream generated by the time-encoding analog-to-digital converter.

21. The micro device according to claim 19, wherein the micro device is arranged to store measured data from the sensor continuously over a certain period of time, and to recall the stored data and apply the data to the load modulation circuit during a period of communication.

22. The micro device according to claim 17, wherein the load modulation circuit comprises at least two electric switches connected to the piezoelectric transducer and arranged for being controlled to modulate electric load of the piezoelectric transducer in response to the one-bit data stream.

23. The micro device according to claim 17, wherein the load modulation circuit is arranged to control electric load of the piezoelectric transducer between a first load state and a second load state in response to the one-bit data stream, wherein the at least one electric switch is controlled so as to provide different electric loads of the piezoelectric transducer in the first load state than in the second load state.

24. The micro device according to claim 23, wherein in the first load state the at least one electric switch is controlled to short-circuit the terminals of the piezoelectric transducer.

25. The micro device according to claim 23, wherein in the second load state the at least one electric switch is controlled to provide an electric load of the terminals of the piezoelectric transducer to cause a minimal backscattering from the piezoelectric transducer.

26. The micro device according to claim 17, wherein the time-encoding analog-to-digital converter is a delta-sigma modulator.

27. The micro device according to claim 17, wherein the sensor is one of: a neural activity sensor such as a Local Field Potential sensor or a single cell sensor, a bio-chemical sensor, a temperature sensor, or a pressure sensor.

28. The micro device according to claim 17, being configured for implantation into brain tissue.

29. The micro device according to claim 17, having a total volume of less than 1 mm.sup.3, such as less than 0.5 mm.sup.3, such as less than 0.2 mm.sup.3.

30. A sensor system comprising: a micro device according to claim 17, an ultrasonic transmitter arranged to transmit an ultrasonic power signal to the micro device; and an ultrasonic receiver arranged to receive the backscattered signal from the piezoelectric transducer of the micro device, and to de-modulate the backscattered signal, to arrive at a representation of a time sequence of the physical parameter measured by the sensor in the micro device.

31. The sensor system according to claim 30, comprising a plurality of micro devices, wherein the sensor in each of the plurality of micro devices comprises a neural activity sensor, and wherein the ultrasonic receiver is arranged to receive backscattered signals from the plurality of micro devices, and to de-modulate the backscattered signals to arrive at representations of respective time sequences of neural activities measured by the plurality of micro devices.

32. A method for transmitting sensor data from a micro device implanted in biological tissue, the method comprises: providing a micro device comprising a piezoelectric transducer connected to a power management circuit for powering power consuming components of the micro device by means of an ultrasonic power signal received by the piezoelectric transducer from an external source, the micro device further comprising a sensor arranged to measure a physical parameter and to generate an electric signal accordingly; digitizing the electric signal from the sensor to generate a one-bit data stream being a representation of the electric signal from the sensor; and modulating electric load of the piezoelectric transducer in response to the one-bit data stream, so that a backscattered signal from the piezoelectric transducer is modulated by the one-bit data stream.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0051] The invention will now be described in more detail with regard to the accompanying figures. The figures show one way of implementing the present invention and is not to be construed as being limiting to other possible embodiments falling within the scope of the attached claim set.

[0052] FIG. 1a illustrates a block diagram of a micro device embodiment, while FIG. 1B illustrates a block diagram of a sensor system embodiment;

[0053] FIG. 2 illustrates steps of a method embodiment;

[0054] FIG. 3 illustrates an example of a load modulation circuit;

[0055] FIG. 4 illustrates graphs showing a simulation of the proposed analog backscattering;

[0056] FIG. 5 illustrates a graph showing acoustic intensity at the piezoelectric transducer in response to a step input of load modulation at the micro device;

[0057] FIG. 6 illustrates a diagram of an example of a SRAM memory with peripherals and interfaces;

[0058] FIG. 7 illustrates an example of a control circuit for the SRAM memory of FIG. 6; and

[0059] FIG. 8 illustrates steps of an example of a design methodology for optimal utilization of an ultrasonic backscattering data channel.

DETAILED DESCRIPTION OF AN EMBODIMENT

[0060] FIG. 1a illustrates a micro device MD, e.g. a dust, embodiment which receives an ultrasonic power signal UPW from an external source by means of a piezoelectric crystal PZC. A power management circuit PMC is connected to electric terminals of the piezoelectric crystal PZC and generates a power output PW accordingly for powering the power consuming components of the micro device MD. A sensor SNS serves to measure a physical parameter related to the biological tissue in which the micro device MD is implanted, e.g. a neural activity signal, such as a Local Field Potential LFP. The sensor SNS generates an electrical signal according to the measured parameter, and this is received by a front end circuit FE which may comprise a pre-amplifier with a high impedance, and optionally further conditioning circuits. The amplified signal from the sensor is then applied to a time-encoding analog-to-digital converter D1B, e.g. a delta-sigma modulator, which generate a one-bit data stream representing the electric signal from the sensor.

[0061] The one-bit data stream is applied to a load modulation circuit LMC with one or more electric switches connected to terminals of the piezoelectric transducer PZC, so as to allow modulation of electric load of the piezoelectric transducer PZC in response to the one-bit data stream, preferably a harsh switch of load to provide a significant load change following the one-bit data stream. Hereby, the piezoelectric transducer PZC will generate a backscattered signal B_U in response to the received ultrasonic power signal UPW which is modulated by the one-bit data stream. Thus, the backscattered ultrasonic signal B_U from the micro device MD contains data generated by the sensor SNS, and an external ultrasonic receiver can pick up this signal B_U and perform a de-modulation to arrive at data representing a time signal sensed by the sensor SNS in the micro device MD. In this way it has been found that a high data rate can be obtained in the transmission of data from the micro device MD. Further, this way of transmitting data only requires a minimum of extra components in the micro device MD, e.g. a delta-sigma modulator, while the piezoelectric crystal PZC component is used for power harvesting as well as data transmission. Thus, with the proposed method for data transmission a high data rate is combined with a low power consumption and a low volume required for extra components.

[0062] It is so be understood that the micro device MD may comprise further components with other functionalities, e.g. one or more light sources (micro LEDs) and further sensors. A wireless receiver may be used to receive wireless control signals for controlling such other components. E.g. for controlling light sources for generating light to the biological tissue for optogenetics and/or, and/or for optically triggering of drug delivery to surrounding biological tissue by providing light on a drug container inside or outside the micro device MD.

[0063] FIG. 1B illustrates a block diagram of a sensor system embodiment for use as a computer-brain interface The dashed box indicates components implanted inside the skull of a person. In this embodiment, a three layer approach is provided for communication and powering of two implantable brain dusts MD1, MD2 which each has a piezoelectric transducer and a sensor. The sensors sense respective neural activity time signals S1, S2. Both ultrasonic power signal UPW and generate respective backscattered ultrasonic signals B_U1, B_U2 in which representations of the sensed neural activity time signals S1, S2 are contained according to the described analog modulation principle.

[0064] A computer CMP outside a person's body is an end-receiver of the neural activity time signal data S1, S2 sensed by the micro device MD1, MD2. A first interface part IF1 is arranged for position outside the skull, on the head, of a person, and this first interface part IF1 serves as interface between the computer CMP and a second interface part IF2. The second interface part IF2 is arranged for implantation inside the skull of the person and serves as interface between the first interface part IF1 and the micro devices MD1, MD2. The second interface part IF2 has an ultrasonic transmitter which transmits the power signal UPW for powering the micro devices MD through the brain tissue. Further, the second interface part IF2 comprise an interrogator part arranged to receive the backscattered ultrasonic signals B_U1, B_U2 through the brain tissue from the micro device MD1, MD2. The first and second interface parts IF1, IF2 are connected by a wireless transmission, e.g. an ultrasonic and/or a radio frequency (RF). A de-modulation processing, including a low-pass filtering, to de-modulate the received backscattered signals B_U1, B_U2, may be performed in the computer CMP, the first interface part IF1 or the second interface part IF2, preferably in the first or second interface parts.

[0065] The first and second interface parts IF, IF2 may further be arranged to communicate wireless control signals for controlling function of the micro devices MD1, MD2, e.g. via ultrasonic signal or via electromagnetic RF signals. Such functions of the micro devices MD1, MD2 may be stimulation of the brain tissue by means of optical and/or electrical signals and/or drug to be controllably released by the micro device. In this way, a computer-brain interface can be implemented, e.g. as a closed-loop control system controlled by a control algorithm in the CMP.

[0066] FIG. 2 illustrates steps of an embodiment of a method for transmitting sensor data from a micro device, e.g. a dust, implanted in biological tissue, e.g. brain tissue. First, providing P_MD a micro device (a dust) comprising a piezoelectric transducer connected to a power management circuit for powering power consuming components of the micro device by means of an ultrasonic power signal received by the piezoelectric transducer from an external source, the micro device further comprising a sensor arranged to measure a physical parameter and to generate an electric signal accordingly. Next, digitizing D_ESS the electric signal from the sensor to generate a one-bit data stream being a representation of the electric signal from the sensor. Next, storing S_TS_1B a time sequence of the one-bit data stream in a memory in the micro device, e.g. a time sequence having a length of 1-100 ms in case of a neural activity sensor. Next, applying A_TS_CR the stored time sequence of one-bit data stream to a load modulation circuit which finally performs modulating M_EL_PZ electric load of the piezoelectric transducer in response to the applied one-bit data stream. Hereby, a resulting backscattered signal from the piezoelectric transducer is modulated by the one-bit data stream. Thus, it is to be understood that the method preferably comprises the step of receiving an ultrasonic power signal by the piezoelectric transducer from an external ultrasonic power source, thus receiving an ultrasonic signal which is then backscattered in a modulated version to represent data indicative of the signal generated by the sensor in the micro device.

[0067] FIG. 3 illustrates an example of a load modulation circuit. An analog signal AS from a sensor is applied to a delta-sigma modulator DSM which generates a 1-bit data stream output controlling a switch arrangement with electric switches M3 and M4 connected to electrical terminals of the piezoelectric crystal PZC for switching its electric load in response to the 1-bit data stream. M3 and M4 are the load modulation circuit in FIG. 3.

[0068] In essence, this circuit modulates the piezoelectric crystal's PZC electrical load harshly according to each bit of the digitized 1-bit data stream by connecting it to either: [0069] 1) an optimum load, i.e. the load that results a minimum ultrasonic reflection from the dust, or [0070] 2) zero-impedance load, i.e. shortening of the crystal's terminal that results in maximum ultrasonic reflection.

[0071] In this circuit, the transistor M3 and M4 in conjunction with active diodes D1 and D2 are building blocks of an active rectifier that converts the AC signal at the terminals of the PZC to a DC voltage for the power consuming elements modeled by a resistive load R.sub.L. C.sub.St is the storage capacitor at the output of the rectifier. The rectifier followed by its loads should results in an optimum load leading to a minimum reflection from the PZC.

[0072] With the proposed modulation, the push of backscattered signal has a low—pass response to the proposed modulations and filters out most of the quantization noise added by the delta-sigma modulator DSM and results in appearance of the analog signal over the push of the backscattered waves. Since, the load and consequently the backscattered signals are modulated harshly in this approach, the amplitude of the modulates signal will be maximized that can results in higher signal to noise ratio of the signal at the interrogator.

[0073] FIG. 4 illustrates graphs showing a simulation of the proposed analog backscattering, namely three graphs indicating amplitude versus time. A verilog-A model for modeling a first-order delta-sigma modulator is used.

[0074] The bottom graph shown an analog 10 kbit sinusoidal input signal, as an example of a sensor analog electric signal.

[0075] A 2 MHz clock to the delta-sigma modulator has been used for sampling, and the middle graph shows the digital output of the delta-sigma modulator.

[0076] The output of the delta-sigma modulator block was fed into a the load modulation circuit (FIG. 3), and the simulated result for the modulated backscattered signal is shown in the upper graph. As seen, the analog backscattering modulation is easily seen, and it has been found that it is possible to de-modulate at the interrogator side, including a low-pass filtering, to arrive at a representation of the original analog input signal with an acceptable signal-to-noise ratio.

[0077] In order to be able to utilize an increase data rate which is possible with the proposed analog modulation backscattering approach for low-bandwidth signals, it can preferably be combined with time compression of the data representing time signal from the sensor, and still allow transmission of the time signal data in real time or at least near real time. This can be obtained by time compression. This can be obtained by storing a time sequence of the sensor output in a 1-bit representation, e.g. an LFP signal having a bandwidth of about 100 Hz, which has been sampled with a reasonable oversampling rate, e.g. 128. By applying recalling the stored 1-bit data stream and applying it to the load modulation circuit at an increased data rate, e.g. with much higher frequency e.g. 100-200 times, a time compression is obtained, and thus e.g. 10 ms of time signal can be transmitted in the backscattered representation in less than 100 μs. Thus, the data capacity can be utilized to transmit time data from additional sensors in a micro device, or it can be used as silent periods to allow other micro device to transmit data over the same ultrasonic channel without interference.

[0078] FIG. 5 illustrates a diagram of an example of a SRAM memory with peripherals and interfaces for implementing such time compression. In the diagram W-CLK and R_CLK are the clock signals that generate timings for writing and reading to the SRAM, respectively. There is one dedicated counter for addressing the SRAM cell for each of the writing and reading operations. Both of the counters get reseat using Power-On-Reset (POR) signal when the micro device, or dust, gets enough power to start working. An 8-bit two-to-one multiplexer is responsible for connecting one of the aforementioned addresses to the SRAM. The SRAM Write and SRAM Read blocks are designed to write or read from the SRAM cells, respectively. Finally, an SRAM Controller generates the controlling signals for writing/reading to/from SRAM.

[0079] One problem that needs to be addressed here is the fact that writing and reading events are not synchronized and they may happen at the same time. Writing in SRAM should be done after each sampling by the sigma-delta modulator. So, for LFP signal with a bandwidth of such as 100 Hz and with an oversampling rate of 128, writing to the SRAM is performed with a frequency of 12.8 kHz and a period of 78.125 μs. On the other hand, reading from the SRAM occurs with the same frequency as the ultrasonic power carrier, e.g. around 3 MHz. If it is intended to receive the backscattered signal in a silent time slot, the duration of a power burst should be equal to two Time of Flight (ToF) of ultrasonic waves in the brain tissue from the UPIB to the dust, which for example is about 40 μs for a distance of 3 cm. Thus, during the 40 μs of reading and load modulation, a maximum of one writing event may happen (writing events are every 78.125 μs). Furthermore, reading has a higher priority in comparison with a writing operation.

[0080] FIG. 6 shows a control circuit which solves this, namely a control circuit designed to generate output signals of Write Enable (Wr_En), Read Enable (R_En), PCH (Pre-Charge), and Decoders' Enable (Dec_En), according to input signals Write Clock(W_CLK), Read Clock (R_CLK) and Write/Read signal (W/R). The control circuit is formed by logic elements, a D-type flip-flop DFF, and falling edge detectors FA_D.

[0081] The DFF is reset during the writing period (when we do not read from the SRAM). Thus, at the beginning of the reading period, Q at the output of the DFF has a low logic value. Then if during the reading period, a writing event happens (falling edge of W_CLK) then the Q switches to the high logic value and again is reset at the end of the reading period. In this case, the falling edge at the output of the DFF will be detected using a falling edge detector and initiates a writing operation.

[0082] In the following a design methodology is discussed for linking choice of various parameters in the design of a system implementing the proposed analog modulation ultrasonic backscattering for data transmission.

[0083] By applying a step to the load modulation circuits, the bandwidth of the backscattering link can be measured using the transient response and the rising time of the backscattered signal.

[0084] FIG. 7 illustrates graphs showing an example of a step load modulation at the dust and the backscattered signal at the external interrogator, namely in the upper graph acoustic intensity at the piezoelectric transducer versus time in response to a step input of load modulation, shown in the lower graph. The results are simulated results. The measurement points illustrates are M_1: 187 mV@106 μs, M_2: 105 mV@78 μs, and M_3: 180 mV@89 μs.

[0085] As seen, the push of the acoustic intensity at the transducer in response to a step input of load modulation at the dust is similar to a first-order low-pass filter, and thus its time function can be driven as below:

l(t)=I0(1−e.sup.−t/τ)

[0086] Where I(t) is the acoustic intensity at the transducer and I0 is its steady state value without any load modulation at the dust. Thus, the time constant of the link (τ), can be calculated by measuring the rising time of the backscattered push. Rising time (tr) is when I(tr) reaches 90% of its initial value and tr=2.23 τ. Thus, based on the simulation in FIG. 7, the time constant and the 3 dB link bandwidth LBW can be derived as below:

[00001]$LBW=\frac{1}{2\pi \tau }$

[0087] In an experiment, the ultrasonic link can be characterized in a similar way. If the bandwidth of the target signal for measurement at the dust is denoted SBW. Then, for time encoding of the data, i.e. converting the data to a 1-bit data stream, sampling with a higher rate than Nyquist rate is required. Over-sampling rate (OSR) of the data has a direct influence in the Signal to Noise Ratio (SNR) of the digitized data. In this way, the sampling rate of the data is given by DSR=2×OSR×SBW.

[0088] It is proposed to store the 1-bit data stream with the rate of DSR. Storing a measured data with time-duration of Tstr, requires a memory with a capacity of at least DSR×Tstr. By using the noise shaping technique, the quantization noise will be moved to the higher frequencies close to DSR/2. However, still the data is in the frequency band of SBW. For taking advantage of the whole link bandwidth efficiently, it is proposed to compress the signal in time domain which is equivalent to expanding the signal in frequency domain. Therefore, the maximum Compression Ratio (CR) to expand the data, excluding the quantization noise, over the whole link bandwidth is CR=LBW/SBW.

[0089] In this way, the noise bandwidth will be increased with the same ratio but out of LBW. Thus, the noise can be attenuated by the link. Attenuation of the quantization noise over the backscattered signal will lead to appearing the analog waveform of the signal over the push of the backscattered signal. The quantization noise can be further removed at the receiver for improving the SNR.

[0090] Furthermore, the proposed time compression is beneficial since the whole stored data can be transmitted in a shorter time and consequently by consuming less energy. Therefore, if the duration of the recorded data is Tstr, the compressed data duration will be Tcmp=Tstr/CR. For doing the time compression with the ratio CR, the data is read from the memory and applied to the load modulation circuit, with a clock frequency of fLM=DSR×CR. By taking the proposed approach, the link capacity (C) can be calculated using the Shannon equation:

C=B×Log.sub.2(1+SNR)

[0091] Here, B is the bandwidth of the transmitted signal over the link. As suggested by this relationship, by expanding the bandwidth of the data to the LBW using the proposed time compression, the capacity of the link can be maximized. The SNR of the signal can be improved with a proper OSR and usage of noise shaping. However, increasing the OSR requires a bigger memory, and more advanced noise shaping can increase the complexity of the circuit implementation. Thus, increasing of the link capacity by improving the SNR of the signal comes at the price of higher power consumption and silicon area.

[0092] FIG. 8 illustrates four steps of a possible design methodology for exploiting the full-bandwidth of an ultrasonic back-scattering channel according to the invention: [0093] 1) Measuring M_BW the bandwidth of the backscattered Link (i.e. LBW) by applying a step function to the dust's load and measuring the rising time of the backscattered signal at the external transducer. [0094] 2) Defining D_DSR the data sampling rate (DSR) by designing of the oversampling rate (OSR) and noise shaping circuits according to the SNR requirements. [0095] DSR=2×OSR×SBW, where SBW is the signal bandwidth. [0096] 3) Defining D_TSQ the time duration of the signal that is measured and stored before each transmission cycle (Tstr), and allocating enough memory (i.e. DSR×Tstr) for storing the data. [0097] 4) Determining D_CR_FLM the compression ratio and load modulation frequency will be driven by CR=LBW/SBW and FLM=DSR×CR, respectively.

[0098] To sum up, the invention provides an implantable micro device, or dust, with a piezoelectric transducer connected to a power management circuit which provides an electric power output for powering components of the micro device based on an ultrasonic power signal from an external ultrasonic signal source. Hereby the micro device is powered. A sensor measures a physical parameter, e.g. a neural activity signal, and generates an electric signal, which is digitized by a time-encoding analog-to-digital converter, e.g. a delta-sigma modulator, to generate a one-bit data stream representing the sensed physical parameter. A load modulation circuit with one or more electric switches connected to the piezoelectric transducer serves to modulate electric load of the piezoelectric transducer according to the one-bit data stream, thus causing a backscattered signal from the piezoelectric transducer to be modulated by the sensed physical parameter. Preferably, the piezoelectric transducer's electric load is harshly modulated by connecting it to either an optimum load for minimum reflection, or short-circuiting for maximum reflection according to each bit of the digitized data stream. Such analog modulation of the data collected at the dust over the backscattered signal to an ultrasonic interrogator increases the ultrasonic data rate capacity. This allows e.g. transmission of time-compressed data from the sensor via ultrasonic backscattering.

[0099] Although the present invention has been described in connection with the specified embodiments, it should not be construed as being in any way limited to the presented examples. The scope of the present invention is to be interpreted in the light of the accompanying claim set. In the context of the claims, the terms “comprising” or “comprises” do not exclude other possible elements or steps. Also, the mentioning of references such as “a” or “an” etc. should not be construed as excluding a plurality. The use of reference signs in the claims with respect to elements indicated in the figures shall also not be construed as limiting the scope of the invention. Furthermore, individual features mentioned in different claims, may possibly be advantageously combined, and the mentioning of these features in different claims does not exclude that a combination of features is not possible and advantageous.