Responses to pure tones and linear FM components of the CF — FM biosonar signal by single units in the inferior colliculus of the mustached bat

Summary

• 1. The responses of 682 single-units in the inferior colliculus (IC) of 13 mustached bats (Pteronotus parnellii parnellii) were measured using pure tones (CF), frequency modulations (FM) and pairs of CF — FM signals mimicking the species' biosonar signal, which are stimuli known to be essential to the responses of CF/CF and FM — FM facilitation neurons in auditory cortex.

• 2. Units were arbitrarily classified into ‘reference frequency’ (RF), ‘FM₂’ and ‘Non-echolocation’ (NE) categories according to the relationship of their best frequencies (BF) to the biosonar signal frequencies (Table 1). RF units have high Q_10dB values and are tuned to the reference frequency of each bat, which ranged between 60.73 and 62.73 kHz (Table 2). FM₂ units had BF's between 50 and 60 kHz, while NE units had BF's outside the ranges of the RF and FM₂ classes.

• 3. PST histograms of the responses revealed discharge patterns such as ‘onset’, ‘onset-bursting’ (most common), ‘on-off’, ‘tonic-on’, ‘pauser’, and ‘chopper’ (Fig. 2). Changes in discharge patterns usually resulted from changes in the frequency and/or intensity of the stimuli, most often involving a change from onset-bursting to on-off. Different patterns were also elicited by CF and FM stimuli (Fig. 6).

• 4. Frequency characteristics and thresholds to CF and FM stimuli were measured. RF neurons (Fig. 3) were very sharply tuned with Q_10dB's ranging from 50–360. Most (92%) also responded to FM₂ stimuli, but 78% were significantly more sensitive (>5dB) to CF stimuli, and only 3% had significantly lower thresholds to FM₂ (Table 4). The best initial frequency for FM₂ sweeps in RF units was 65.35+-2.138 kHz (n=118), well above the natural frequency of the 2nd harmonic (Table 2). FM₂ and NE units were indistinguishable from each other, but were quite different from RF units: 41 % of these two classes had lower thresholds to CF, 49% were about equally sensitive, and 10% had lower thresholds to FM. For FM₂ units, mean best initial frequency for FM was 60.94 kHz±3.162 kHz (n=114), which is closely matched to the 2nd harmonic in the biosonar signal. Very few units (5) responded only to FM signals, i.e., were FM-specialized.

• 5. The characteristics of spike-count functions were determined in 587 units (Table 5). The vast majority (79%) of RF units (n=228) were non-monotonic, and about 22% had upper-thresholds. FM₂ units (77=174) were more evenly divided, 46% monotonic and 54% non-monotonic, while 52% of NE units were monotonie, 48% non-monotonic.

• 6. Responses to CF — FM stimuli mimicking the biosonar signal of FM₂ units depended on the unit's particular thresholds and BF's to CF and FM stimuli alone (Fig. 7). During echolocation, FM₂ units could be excited only by FM₂ components in either or both the emitted pulse and returning echo, but some that had high CFBF's (> 57 kHz) would also respond to moderate-to-high intensity CF components of Doppler-shifted echoes.

• 7. In response to 1st or 2nd harmonic pulse-echo pairs of CF — FM stimuli, FM₂ units showed no facilitation at specific time delays like that seen in the FM — FM area of auditory cortex (Fig. 8). For RF units, 2nd harmonic-CF responses were often suppressed by adding the fundamentals, while such CF₁/CF₂ stimulus pairs caused facilitation in the CF₁/CF₂ region of cortex.

• 8. It is concluded that facilitation by pairs of CF harmonics or FM components with specific time delays is a response property of units only at levels higher than the inferior colliculus in the mustached bat. However, the properties of RF, FM₂, and NE units tuned to the 30 kHz-fundamental (‘FM₁’ units) are similar to the frequency and intensity characteristics of cortical facilitation units, and they are likely to subserve the neural circuitry in the MGB or auditory cortex responsible for such facilitation.