Specific long-term memory traces in primary auditory cortex

Abstract

Learning and memory involve the storage of specific sensory experiences. However, until recently the idea that the primary sensory cortices could store specific memory traces had received little attention. Converging evidence obtained using techniques from sensory physiology and the neurobiology of learning and memory supports the idea that the primary auditory cortex acquires and retains specific memory traces about the behavioural significance of selected sounds. The cholinergic system of the nucleus basalis, when properly engaged, is sufficient to induce both specific memory traces and specific behavioural memory. A contemporary view of the primary auditory cortex should incorporate its mnemonic and other cognitive functions.

Figure 1. The effects of learning on the frequency tuning of neurons in A1

a | Receptive field plasticity of a single cell in auditory cortex, showing frequency tuning functions (70 dB) before and after tone-shock conditioning, and the resultant shift in tuning to the frequency of the conditioned stimulus (CS). The inset shows the difference in tuning (post minus pre), with the maximum increase in response at the frequency of the CS. Modified, with permission, from ref. © (1997) Academic Press. b | Normalized group pre-post difference functions, showing change in response as a function of octave distance from the CS frequency. Conditioning (left) produces a specific increase in A1 response to the CS frequency with reduced responses at most frequency distances. Sensitization training produces a non-specific increase in response across all frequencies, both for auditory sensitization (tone-shock unpaired) and visual sensitization (light-shock unpaired), showing that this non-associative effect is transmodal. Repeated presentation of the same tone alone (habituation) produces a specific decreased response at that frequency. Reproduced, with permission, from ref. © (1995) MIT Press. REP, repeated frequency.

Figure 2. Development and retention of specific receptive field (RF) plasticity

a | Rapid induction (five trials) of RF plasticity, shown as vector diagrams of changes in response to the pre-training best frequency (BF) and the conditioned stimulus (CS) frequency, from suprathreshold (75 dB) responses for individual subjects. Left: after five trials, responses to the BF had decreased while those to the CS increased. Changes were maintained after 15 and 30 trials, but further change developed after one hour (consolidation), at which time the CS frequency became the new BF. Right: sign change in which the CS frequency was inhibitory before training but became excitatory after only five trials; the initial response to the CS was too weak for it to become the new BF or to exhibit consolidation in one hour. Reproduced, with permission, from ref. © (1993) American Psychological Association. b | Long-term consolidation (group RF data) in which responses to the CS frequency increased relative to the pre-training BF over three days (72 hours) and then stabilized over ten days. The effects were significantly different from those for an unpaired group that was studied to seven days post-training.

Figure 3. Specific tuning plasticity induced during appetitive conditioning revealed both in suprathreshold receptive field (RF) and threshold map measures

a | Rats received a tone paired with electrical stimulation of positive reinforcement neurons in the medial forebrain bundle. After a single session (30 trials), single neurons in A1 exhibited CS-specific suprathreshold RF plasticity. The graphs show an example of single unit RFs (left) and RF difference (post minus pre-training) (right) for a case in which tuning did not shift because the CS frequency initially elicited little excitation. However, conditioning produced an increased response only to the CS frequency. Left panel modified, with permission, from ref. © (2001) Blackwell Publishing. b | Rats received a tone paired with electrical stimulation of ‘reward’ neurons in the ventral tegmental area (VTA). Shown are threshold maps of frequency representation for a naive and an experimental animal. Pairing produced a marked increase in the area of representation of the paired tone (yellow area). Modified, with permission, from ref. © (2001) Macmillan Magazines Ltd.

Figure 4. Instrumental appetitive training with a normal reinforcer (bar press rewarded by water only in the presence of a 6-kHz tone) produces a specific increased representation that reflects the level of stimulus importance

a | Individual maps of naive (left) and trained subjects (right) and their respective area quantifications (below each); the naive function is the mean of five rats. The trained rat achieved a high level of 91% correct and an expansion of the octave centred on the 6-kHz signal tone from an average naive value of ~12% to ~45%. Shading denotes the CS frequency band. b | Across subjects, the higher the level of performance, the greater the area representing the training frequency. Modified, with permission, from ref. © (2003) Academic Press. CF, characteristic (threshold) frequency.

Figure 5. Two models of CS-specific tuning plasticity in the primary auditory cortex and fear conditioning

a | The model proposed by Weinberger et al.^,. b | The model proposed by Suga and Ma. ACh, acetylcholine; CS, conditioned stimulus; CN, cochlear nucleus; DCN, dorsal column nuclei; IC, inferior colliculus; MGm, medial geniculate body/posterior intralaminar complex; MGv, ventral medial geniculate body; lat. LN, lateral leminiscal nucleus; PIN, posterior intralaminar nucleus; SOC, superior olivary complex; TRN, thalamic reticular nucleus; US, unconditioned stimulus; vent. post. LTN, ventro-postero-lateral thalamic nucleus. See text for detailed discussion of the two models.

Figure 6. CS-specific tuning plasticity induced by pairing a tone with stimulation of the nucleus basalis (NB), the major supplier of cortical acetylcholine

Shown are group normalized difference tuning functions (post-training minus pre-training). Immediately after training, there was a small increase at the conditioned stimulus (CS) frequency (arrow) that became much larger and more specific 20 min later. Twenty-four hours later, this CS-specific effect had increased further. So, properly timed activation of the NB is sufficient to induce associative, specific, long-lasting tuning plasticity that increases in strength in the absence of additional training (that is, develops consolidation). Modified, with permission, from ref. © (1998) American Psychological Association.

Figure 7. Induction of specific behavioural memory by pairing a tone with stimulation of the nucleus basalis (NB)

Rats received either a 6-kHz tone paired with NB stimulation or the two stimuli unpaired. Twenty-four hours after the end of training, behavioural generalization gradients were obtained by presenting the CS and several other frequencies. Left: both interruption of ongoing respiratory rhythm and change in heart rate were maximal at the CS frequency for the paired group. Right: differences between groups (paired minus unpaired) reveal the associative, specific behavioural effect at the CS frequency, indicating the induction of specific behavioural memory. Modified, with permission, from ref. © (2002) National Academy of Sciences USA. RCI, respiration change index.