A CMOS-based Lab-on-Chip Array for the Combined Magnetic Stimulation and Opto-Chemical Sensing of Neural Tissue Timothy G. Constandinou, Pantelis Georgiou, Themistoklis Prodromakis, and Chris Toumazou Institute of Biomedical Engineering, Imperial College London SW7 2AZ, UK Email:{t.constandinou,pantelis,t.prodromakis,c.toumazou}@imperial.ac.uk Abstract—This paper presents a novel CMOS-based lab- on-chip platform for non-contact magnetic stimulation and recording of neural tissue. The proposed system is the first of its kind to integrate magnetic-stimulation and opto- chemical sensing in a single pixel, tesselated to form an 8×8 array. Fabricated in a commercially-available 0.35μm CMOS technology, the system can be intrinsically used for both optical imaging and pH sensing and includes mecha- nisms for calibrating out sensor variation and mismatch. In addition to sensory acquisition via an integrated 10-bit ADC, a 64-instruction spatiotemporal pattern generator has been embedded within the array for driving the micro- scale magnetic neural stimulation. In this application the ISFET-based sensors are used to capacitively-couple neuronal charge in close proximity to the floating gate. Optical imaging hardware has also been embedded to provide topographic detail of the neural tissue. I. I NTRODUCTION In the last decade, the relentless trend towards technol- ogy miniaturisation has facilitated new opportunities for microfluidic integration. Combined with modern micro- electronics, this has opened up new paradigms for novel lab-on-chip platforms, miniaturising conventional analyt- ical systems. As a result there has been a new drive to fabricate these using standard CMOS technologies due to overlapping requirements when combined with modern day chemical micro-sensing devices [1]. These include miniaturisation of sensors, batch fabrication, integration of processing and signal conditioning circuitry and cheap cost of manufacture. These enabling technologies have enabled spatiotem- poral sensing and excitation of various physical and chemical parameters by exploiting parasitic devices. For example, CMOS-based imagers using parasitic pn- junction photodiodes have been developed into novel lab-on-chip platforms for applications in fluorometry to measure metabolic activity and viability of biological cells [2]. Similarly, ISFET-based chemical sensors are becoming increasingly popular due to the fact that they can be fabricated in unmodified CMOS technology. These can sense pH by exploiting the silicon nitride passivation which can be used as an sensing membrane [3], [4]. In [5], an alternative modality for sensing with ISFET devices was reported where neural action poten- tials were sensed through electrostatic coupling between neuron and floating gate. Almost 30 years ago [6], it was demonstrated that magnetic stimulation is possible by aid of eddy cur- rents. When compared with traditional electrical stim- ulation, this method provides certain advantages such as non-invasiveness, improved biocompatibility and bio- resistance matching [7], [8]. The lack of a metal- electrolyte interface, which often posses challenges such as modification of the electrode surface, corrosion and bio-fouling is alleviated, resulting in improved robust- ness and system stability. Recent advancements in tran- scranial magnetic stimulation (TMS) utilize arrays of coils, which offer the added benefits of spatial-temporal in addition to focused magnetic stimulation [9]. In CMOS technology, exploiting thick top metal process options adopted for RF applications, microcoils have also been applied to lab-on-chip platforms [10]. Used in an array, these can generate a spatiotemporal magnetic field to facilitate neural stimulation. We report the first lab-on-chip array for the non- contact interfacing to neural tissue using magnetic stim- ulation and electrostatic sensing, in addition to opti- cal imaging. This is facilitated using inductive micro- coils and ISFET-based charge sensors for the neural interfacing and an Active Pixel Sensor (APS) for the optical imaging. The system is scaled to form an 8×8 array capable of calibrating out sensor non-idealities including drift, spatial (static) noise and gain mismatch while a programmable spatial field generator for focused magnetic stimulation is also provided. 2010 12th International Workshop on Cellular Nanoscale Networks and their Applications (CNNA) 978-1-4244-6678-8/10/$26.00 ©2010 IEEE