Before one can begin talking to nerves and neurons, one must first determine how they communicate with one another and determine how information is encoded within the communication. The challenge, then, is to monitor the signals transmitted by one or more neurons to a tightly controlled natural stimulus and correlate features of the signals with information contained in the stimulus.
Several methods have been devised to record from nerves and neurons, based on biological knowledge of how nerves conduct signals. Most nerves communicate via action potentials, a complex signal generated by an intricate coordination of ion movements across neuronal membranes (see Chapters two and three) and controlled by voltage potentials across the cell membrane. Recording devices must therefore tap or intercept voltages and ionic currents, and transform them into electrical signals suitable for processing. While the concept is relatively simple, implementing such a device is complicated by the millimeter to micrometer scale of most neurons and the small changes (millivolts or lower) in membrane potentials typically encountered. Material scientists are required to develop devices small and reliable enough to interact with a neuron. Devices such as the cuff electrode and suction electrode8 measure a compound signal from the entire nerve, while single-wire electrodes and electrode arrays (see Chapter seven) aim to record from one or a small population of neurons, respectively. Electrical engineering techniques are required to extract the neural signal from biological and external noise sources and amplify them to manageable levels for processing. Solutions are found in physical differential amplifier-based head stages and in analog filters and amplifiers, in addition to the software or microprocessor-based solutions that use sampled representations of the recorded signal.
Correlating features of the neural signal with the original stimulus requires knowledge of the biological system under study and powerful analysis tools to examine the myriad of possibilities. In the ear, for example, the axons that extend from the cochlea to make up the auditory nerve are known to respond to specific sonic frequency ranges distributed along the length of the cochlea (see Chapter ten). This knowledge of how the neural system functions provides important clues to interpreting signals from different sections of the auditory nerve. It is also generally agreed upon that the frequency, timing, and duration of action potentials generated by a given neuron carry a significant amount of information. Analysis tools must therefore be able to track amplitude and frequency as a function of time. Wavelet theory has proven to be a useful tool, though not the only one, in addressing this challenge. Artificial neural networks and other sophisticated statistical analysis programs provide the computational power necessary to sift through a variety of possibilities and to arrive at the most likely relationships between stimulus and response (see Chapter twelve for example).
Was this article helpful?