The suboesophageal ganglion: a ‘missing link’ in the auditory pathway of the locust

Summary

• 1. Five interneurones that receive auditory input are described in the Suboesophageal ganglion [SOG] of the locustLocusta migratoria [Fig. 1]. The neurones have been characterised physiologically and anatomically by intracellular recording and staining with Lucifer Yellow.

• 2. The newly described neurones all have their somata in the SOG (Fig. 1). They are classified into three groups: (a) ascending neurones with axons in the circumoesophageal connectives (SA1, 2, 3); (b) a local neurone with no axon (SL1); and (c) a descending neurone with its axon in the cervical connective (SD1).

• 3. Transverse sections of the labial and maxillary neuromeres show that the arborisations of the SOG auditory neurones occupy two main areas, medio-dorsal and medio-ventral (Figs. 2, 3). Two auditory interneurones that ascend from the thorax to the brain, the G and B neurones, have segmentally arranged axon collaterals in the SOG (Fig. 1) which project into the same areas of neuropil.

• 4. In the brain, SA1 terminates in the same dorsal areas of the lower lateral lobe of the protocerebrum as the thoracic ascending auditory interneurones (Fig. 4). SA2 and SA3 terminate more medially where local and descending interneurones arborise.

• 5. The SOG neurones show sigmoid intensity-response curves (Fig. 7) and broad-band frequency-response curves with maxima between 10 and 20 kHz (Fig. 6), similar to those of the thoracic ascenders at comparable intensities. The SOG neurones have spiking thresholds for auditory stimuli somewhat higher than those of thoracic ascending neurones.

• 6. All neurones except SD1 spike in response to tones. Discharges are phasic or phasic/tonic in response to short (20 ms) tones, but can become tonic when stimulated with longer tones and depending on the extent of habituation. The SD1 spiked in response to sound signals only with concomitant depolarisation by current injection (Fig. 8B). The oscillating nature of the responses to tones in SD1 (Fig. 8 Biv) and SA2 (Figs. 10B, 11) suggest a recurrent pathway, perhaps involving axon collaterals.

• 7. Auditory information is conducted from the thorax to the SOG as spikes, not as passively conducted potentials (Fig. 5). Spike latencies in the SOG neurones are long (37–85 ms) compared with those of thoracic ascenders recorded in the SOG (Table 1). In SL1, however, the delay between the spike in a thoracic ascender and the first stimulusrelated EPSP is much shorter (ca. 1 ms) making it a likely candidate to receive information directly from through-fibres.

• 8. In SA2, SA3, SL1 and SD1, light stimuli also produce EPSPs (Fig. 10) which are not influenced by EPSPs to sound because responses to light and to sound decrement independently. The visual input to ascending SOG neurones indicates there are information loops between the brain and SOG.

• 9. The responses of SA2, SA3, SL1 and SD1, but not SA1, decrement to varying degrees on repetitive stimulation. The SA3 neurone, which has axons in both circumoesophageal connectives (Fig. 1), displays side-dependent habituation.

• 10. The neurones described here are components of a previously unknown auditory processing centre in the locust CNS. Their anatomical and physiological properties suggest that the role of the SOG in the auditory pathway needs reassessing, and the concept of the locust auditory system as a linear hierarchy may no longer be valid.