Class warfare resolved in the auditory midbrain

Abstract

Neurophysiologists are human, and biology is complex. To deal with this complexity, nature endowed us with the tendency to organize complex populations into categories. Class distinctions are a hallmark of how we deal with society. We draw distinctions between peoples with different languages, accents, or economic status. Our subjective categories or classifications can be quite useful in helping bring order to our world. Class distinctions also are a hallmark of how neuroscience deals with the complexity of the nervous system. Some may consider neurons to be just one type of cell in the body, but in reality the population of neurons is complex. There are many variations of neurons that differ along many dimensions: morphology, intrinsic membrane properties, composition and location of synapses, or the role in a neural network. Electrophysiologists who study neurons in networks or systems have an especially challenging task since the activity of neurons is observed while they process different types of information. This activity changes in response to different types of information and may be dynamic. It is natural for the physiologist to subjectively classify neural responses. What may be unusual is to question the validity of those classifications. Yet, that is precisely the approach of the paper in this issue of The Journal of Physiology by Palmer et al. (2013), Classification of frequency response areas in the inferior colliculus reveals continua not discrete classes. The auditory system is a very complex neural network (Fig. 1). Compared to somatosensory and visual systems, whose information jumps to the cortex through just one synapse in the thalamus, information from the ear may pass through three or more separate processing stages before arriving in auditory cortex. The information emerging from the cochlea via the eighth cranial nerve is remarkably uniform. However, when the auditory nerve branches at the cochlear nucleus, it terminates on different neuron types. In turn, these initiate different auditory brainstem pathways that synapse in distinct brainstem regions, each performing unique neural computations and extracting different types of information about sound. These parallel information-processing channels converge in the inferior colliculus (IC) of the midbrain where recordings show a wide diversity of neural responses to sound. Anatomical evidence (Fig. 1) supports the idea that IC neurons with different response properties to sound may receive only a subset of the many ascending excitatory and inhibitory inputs to the IC (Loftus et al. 2010). The natural reaction of auditory physiology is to classify these responses of IC neurons to sound. Figure 1 Auditory pathways leading to and from the inferior colliculus The response to sound frequency and intensity is one of the primary hallmarks of a neuron in the auditory system. The spectral response across many sound frequencies and intensities, frequency response area, is a basic metric of all auditory neurons. Numerous studies of the IC suggest there are different classes of spectral responses. However, these are subjective classifications by investigators who often disagree on which ones are the ‘proper’ classifications. One could say there was even ‘class warfare’ as each camp of physiologists favours its own classification scheme. Here, Palmer et al. use a different approach to address the issue of the classification of frequency response areas in the IC. Two things are remarkable about this study. First, they have assembled literally a lifetime's supply of data on the frequency responses of IC neurons in the guinea pig – thousands of neurons, recorded using basically the same methodology with colleagues over the years. Since this must surely be the largest sample of IC neurons to be analysed at one time, it should undoubtably sort out the types of frequency responses. Of course, this data set includes exemplars of frequency responses used to subjectively classify IC neurons. However, the second thing and real innovation here is the application of an objective methodology to sort the data. In doing so, a different picture emerges about the frequency coding of neurons in the IC. It is more complex that we wish it to be. While examples of most, but not all, categories recognized by the subjective classification are seen, these are merely the peaks of the mountains in a mountain range, and there is a great deal happening between the peaks. A tremendous amount of integration of both excitatory and inhibitory information occurs in the frequency dimension. Thus, the lack of separable, discrete classes for frequency coding provides a new insight into the biology of the auditory midbrain. It also shows how we sometimes create more contrast between the categories than may exist in real life.