Sirs

These comments relate to the recent review by Weinberger1. In the big brown bat, 'tone-specific' best frequency (BF) shifts in the auditory cortex (AC) and the inferior colliculus are evoked by neither acoustic stimulation (CS) alone, nor leg-stimulation (US) alone, but by paired stimulation (CS–US), that is, by auditory fear conditioning. Backward conditioning (US–CS) does not evoke these BF shifts. The BF shifts are specific to the frequency of the conditioned tone (the CS). Unlike the collicular BF shift, the cortical BF shift increases and reaches a plateau after the termination of conditioning. This plateau is sustained for a long period of time2,3,4. This long-term cortical BF shift is elicited by an increase in cortical acetylcholine3,4,5. Therefore, the cortical BF shift represents physiological associative memory.

A neural network for evoking the long-term cortical BF shift is proposed by Suga and co-workers2,3. Our hypothesis states that small and short-term tone-specific cortical and collicular BF shifts are evoked by the AC, and corticofugal feedback activated by tone stimulation alone, and that the cortical BF shift is augmented and changed into a long-term shift by increased cortical acetylcholine elicited by the basal forbrain, which is activated by the auditory and somatosensory cortices through the amygdala. This hypothesis, or model, differs from Weinberger's model6, in which CS–US association occurs in the multi-sensory thalamic nuclei (magnocellular division of the medial geniculate body/posterior intralaminar nucleus (MGBm/PIN)), the MGBm–AC projection evokes a small short-term cortical BF shift, and this BF shift is augmented and changed into a long-term shift by an increase in cortical acetylecholine elicited by the basal forebrain, which is activated by the MGBm/PIN through the amygdala.

The Gao-Suga hypothesis is based on the following neurophysiological data7.

(i) Tone-specific cortical and collicular BF shifts can be evoked by either electric stimulation of the AC alone7,8,9,10,11 or by 30-min-long repetitive acoustic stimulation alone (not the short trains of acoustic stimulation used as a CS)9,12,13. These BF shifts are evoked by a mechanism that is 'intrinsic' to the auditory system, without CS–US association in the MGBm/PIN. These findings do not contradict any existing neurophysiological data because neither electric stimulation nor lesion experiments have been carried out to show that the MGBm-to-AC projection evokes the tone-specific cortical BF shift.

(ii) Muscimol applied to the AC blocks the development of the conditioning-dependent short-term collicular BF shift12. Atropine applied to the AC blocks the development of the conditioning-dependent long-term cortical BF shift, but not the collicular BF shift4. Atropine applied to the inferior colliculus abolishes the development of the collicular BF shift, reduces the cortical BF shift and makes it short-term4. Therefore, the corticofugal feedback evokes the short-term collicular BF shift, and this collicular BF shift contributes to the large, long-term cortical BF shift.

(iii) Inactivation of the somatosensory cortex by muscimol does not affect cortical auditory responses, but selectively abolishes the development of the conditioning-dependent cortical and collicular BF shifts2. Electric stimulation of the somatosensory cortex after, but not before, acoustic stimulation or electric stimulation of the AC augments the cortical and collicular BF shifts7,9. This augmentation does not occur if the basal forebrain is lesioned7. Therefore the somatosensory cortex, through the cholinergic basal forebrain, has an essential role in the development of conditioning-dependent BF shifts.

(iv) Inactivation of the amygdala prevents the development of conditioning-dependent plastic changes in the MGBm/PIN, although there is no backward connection from the amygdala to the MGBm/PIN14,15. Therefore, the MGBm/PIN is not the first place where CS–US association occurs.