MEA Technology
Multi Electrode Arrays (MEAs) represent a widely used methodology to access electrophysiological signals in in-vivo or in-vitro electrogenic cell networks. It is a unique non invasive (i.e. extracellular) approach, that allows long-term monitoring of large neuronal cell assemblies to investigate fundamental properties (e.g. plasticity, memory) of networks or to assess pharmacologically induced responses or neurotoxicity. However, conventional MEA systems implement up to 256 electrodes and provide only a partial view of the network signaling. These MEA-chips are microfabricated devices that embed an array of metallic microelectrodes on a glass or silicon substrate (carrier) with a typical microelectrode diameter in the range of 10µm-30µm and with a center-to-center pitch of more than 100µm. Each microelectrode is individually wired to a bond pad at the edge of the chip and thus it is not possible to reach dense electrode arrays over large active areas. The active CMOS technology developed by 3Brain allows to overcome this limitation and to implement large and dense multi electrode arrays to map extracellular signaling in cultures and tissues.
Today’s Neuroscience research needs adequate electrophysiological instrumentation to scale up recordings from single cells to large networks. Moreover, detailed information from each experiment is expected in order to maximize the scientific outcomes and reduce costs. The trend is towards more data points per experiment to measure electrophysiological events from the cellular scale to the network scale. Simple microfabrication technologies, such as the ones used for conventional MEAs, are not sufficient to comply with the increasing need of more electrodes. Routing and interconnection complexity limit conventional MEAs from scaling up the number of electorde channels to more than a few hundreds and a new technological approach is needed for next-generation devices implementing several thousands of channels.
Complementary Metal Oxide Semiconductor (CMOS) technology can bridge the gap between low-electrode-counts MEA devices and high-resolution MEA platforms. CMOS allows cost-efficient implementation of electrodes and electronics onto one single silicon chips. No more complex interconnection is required to take low amplitude signals from the electrodes to an external bank of amplifiers. The amplifiers are directly integrated underneath each electrode and therefore this hi-tech approach allows the implementation of MEA devices with thousands of electrodes without increasing the interconnection complexity. On-chip switching (i.e. multiplexing) enables to take hundreds of electrode channels on one single output line.
3Brian high resolution MEA systems are based on proprietary circuits integrated in CMOS technology (figure 1). Each electrode is integrated in a “pixel” and a very high degree of integration is provided by locating an amplifier/filter-circuitry underneath each individual electrode. This concept originally stems from the technology of CMOS-based optical cameras, where the light sensitive elements (i.e. photo diodes) are replaced with metallic electrodes. An element consisting of a metallic electrode and local electronics for signal conditioning is referred here as a “pixel” (figure 2). Each pixel’s output is switched to one single output at a defined time. In this way, each pixel within an array is read one after each other. They all have a precise timing relationships that allows decreasing the number of effective output channels. Therefore, one output channel contains the multiplexed signals of many electrode channels and will be re-ordered by the BioCam hardware. The concept of figure 2 is borrowed from the optical imager field. It is called APS-Technology (Active Pixel Sensor Technology). APS means that active signal processing takes place locally at the electrode level.
 |
|
 |
| Figure 1: CMOS microchip |
|
Figure 2 Array of Pixel Elements. Each pixel consists of an amplifier and an electrode is arbitrarily addressable. |
Real-Time Technology
Since thousands of recording electrodes generate a large amount of data, 3Brian developed real-time solutions to manage this data flow. The sampling rate of one electrode has to be at several kHz to resolve spiking activity, and therefore the total data rate resulting from the system can easily go up to 500 kb/s. Thus, the platform requires efficient and fast data processing units provided with the BioCAM platform and BrainWave software. Briefly, the data stream is pre-processed in a Field Programmable Gate Array (FPGA). FPGAs allow the implementation of time-efficient hardware processing units. As opposite to digital signal processors (DSPs) that can be programmed in software and that are also efficient for specific classes of operations (i.e. additions, multiplications), FPGA implementations can map many common tasks in a specific and optimized set of hardware-blocks (digital logic). This approach is the fastest way to reliably process data.
Our system architecture uses both hardware- and software-level signal processing to optimize speed and flexibility of algorithms. Signal conditioning is done in hardware and data is subsequently processed in software, designed to exploit multicore environments. Based on this approach the optimized 3Brain hardware and software allow on-line visualization and spike detection of the full array of 4096 electrodes while running your experiments.The full array signals are visualized as colour coded movies, but you are also free to visualize multiple electrodes as voltage-time plots.
Figure 3: Architecture of BioCAM 4096.