A new micromachining process for batch fabrication of silicon neural probes has been developed and demonstrated in the design and manufacture of probes with 32- and 64-site recording electrode arrays.
It is believed that a key to the understanding of the neural system is to make simultaneous observations of the activity of a large number of cells. Probes that can penetrate neural tissue and insert a large number of recording sites, while tissue damage is kept to a minimum, are thus needed, and microsystem technology (MST) is well suited to this end. In comparison, conventional wire electrode technology is limited to fewer recording sites in a small volume of tissue, has less exact relative positioning and/or will cause more damage to the neural tissue.
Although micro-machined neural probes have been shown previously, the most commonly used process is based on a combination of wet silicon etching and a p++ diffused etch stop to create fork-like probe shapes. In contrast, we have developed an all-dry silicon-etch based process where the buried oxide layer of a silicon-on-insulator (SOI) substrate acts as an etch stop for a double-sided deep reactive ion etch (DRIE). We believe that this process has several attractive features when considering process related design limitations, process complexity, uniformity and manufacturability. Double-sided lithography and etching allow a thicker base plate to be part of the design, which facilitates the handling of the otherwise very thin probes after processing. The use of step-and-repeat projection lithography enabled structures with down to 1 um line-width. This gives a better trade-off between electrode-count and tissue damage than most other probes available today.
The probe chips are mounted on custom designed flexible printed circuits (FPC). Electrical interconnects are made with ultrasonic wire bonding. The chips and bond wires are protected with an epoxy glob top. The back end of the FPC is designed to mate with a zero insertion force (ZIF) connector.
Measurement and testing contribute a significant part to the cost of microsystems. Automated measurement set-ups are difficult to arrange for microsensors, which must interact with the environment during the test. In this project we have also demonstrated a novel 3-point impedance measurement procedure for microelectrodes. The concept is based on the use of a saline gel and enables relatively fast measurement at wafer level.
a) Basic neuroscience research:
Recordings and correlation analysis of multiple signals are made on animal models (e.g. rat, mouse, and guinea pig) to improve basic knowledge of the function of the brain and neural disorders. This application area also includes research on computer algorithms modelled on the function of the brain. A market estimate, based on the number of relevant presentations during the Society for Neurosciences Annual Meetings or the number of its relevant members, leads to the conclusion, that there are 2000 laboratories in North-America engaged in electrophysiological research and therefore potentially interested in acquiring a multi-site recording system. There is today no known commercial competitor producing this type of fork-like neural probes, although academic groups have provided U.S. neuroscientists with probes on a non-commercial basis.
b) Pharmaceutical research and drug development:
Part of the development cost of a new drug is spent in order to assess the activity of the new drugs in models. However, in particular neurologic active drugs for humans should not be tested just in cell cultures. Due to the unpredictable influence of the drug on higher-level neuronal networks, they have to be tested electrophysiologically on animals. Being able to perform several dozens instead of a handful of recordings at the same time with the same animal and perfect control of the recording sites to each other would mean a significant improvement in testing performance. Savings on the number of lab animals could be realised, not to mention the invaluable data on inter-neuron correlation. An estimate of this market segment leads to at least double the size of the basic research segment.
c) Medical treatment, e.g. targeting of brain structures during brain surgery:
d) Interfaces for control of prostheses (brain-machine interfaces, BMIs:
This two last application areas may require new probe materials.