Single-cell activity was recorded in the dorsal subnucleus of the lateral amygdala (LAd) of freely behaving rats during Pavlovian fear conditioning, to determine the relationship between neuronal activity and behavioral learning. Neuronal responses elicited by the conditioned stimulus typically increased before behavioral fear was evident, supporting the hypothesis that neural changes in LAd account for the conditioning of behavior. Furthermore, two types of these rapidly modified cells were found. Some, located in the dorsal tip of LAd, exhibited short-latency responses (<20 ms) that were only transiently changed. A second class of cells, most commonly found in ventral regions of LAd, had longer latency responses, but maintained enhanced responding throughout training and even through extinction. These anatomically distinct cells in LAd may be differentially involved in the initiation of learning and long-term memory storage.

The GABAa agonist, muscimol (0.5 jig in 0.5 ul saline), or vehicle was infused into the lateral and basal amygdala nuclei prior to fear conditioning or testing in rats. Rats given muscimol before conditioning and saline before testing showed much less freezing to the conditioned stimulus (CS) and the context than did controls given saline before training and testing. Rats given saline before training and muscimol prior to testing also showed low levels of freezing to the CS and the context. In follow-up procedures, rats with acquisition initially blocked by pretraining muscimol infusions froze in a manner similar to that of controls when retrained and retested with saline infusions. Rats trained with saline but tested with muscimol presumably became conditioned but could express the learning. When retested with saline, they froze in the same manner as controls. Thus, activity in the lateral and basal amygdala appears to play an essential role in the acquisition and expression of fear conditioning.

Long-term potentiation (LTP) and long-term depression (LTD) underlie most models of learning and memory, but neural activity would grow or shrink in an uncontrolled manner, if not guarded by stabilizing mechanisms. The Bienenstock-Cooper-Munro (BCM) rule proposes a sliding threshold for LTP/LTD induction: LTP induction becomes more difficult if neural activity was high previously. Here we tested if this form of homeostatic plasticity applies to the human motor cortex (M1) in vivo by examining the interactions between two consecutive sessions of paired associative stimulation (PAS). PAS consisted of repeated pairs of electrical stimulation of the right median nerve followed by transcranial magnetic stimulation of the left M1. The first PAS session employed an interstimulus interval equalling the individual N20-latency of the median nerve somatosensory-evoked cortical potential plus 2 ms, N20-latency minus 5 ms, or a random alternation between these intervals, to induce an LTP-like increase in motor-evoked potential (MEP) amplitudes in the right abductor pollicis brevis muscle (PAS(LTP)), an LTD-like decrease (PAS(LTD)), or no change (PAS(Control)), respectively. The second PAS session 30 min later was always PAS(LTP). It induced an moderate LTP-like effect if conditioned by PAS(Control), which increased if conditioned by PAS(LTD), but decreased if conditioned by PAS(LTP). Effects on MEP amplitude induced by the second PAS session exhibited a negative linear correlation with those in the first PAS session. Because the two PAS sessions activate identical neuronal circuits, we conclude that 'homosynaptic-like' homeostatic mechanisms in accord with the BCM rule contribute to regulating plasticity in human M1.

Coincident hand and foot movements are more reliably performed in the same direction than in opposite directions. Using transcranial magnetic stimulation (TMS) to assess motor cortex function, we examined the physiological basis of these movements across three novel experiments. Experiment 1 demonstrated that upper limb corticomotor excitability changed in a way that facilitated isodirectional movements of the hand and foot, during phasic and isometric muscle activation conditions. Experiment 2 demonstrated that motor cortex inhibition was modified with active, but not passive, foot movement in a manner that facilitated hand movement in the direction of foot movement. Together, these findings demonstrate that the coupling between motor representations within motor cortex is activity dependent. Because there are no known connections between hand and foot areas within primary motor cortex, experiment 3 used a dual-coil paired-pulse TMS protocol to examine functional connectivity between secondary and primary motor areas during active ankle dorsiflexion and plantarflexion. Dorsal premotor cortex (PMd) and supplementary motor area (SMA) conditioning, but not ventral premotor cortex (PMv) conditioning, produced distinct phases of task-dependent modulation of excitability of forearm representations within primary motor cortex (M1). Networks involving PMd-M1 facilitate isodirectional movements of hand and foot, whereas networks involving SMA-M1 facilitate corticomotor pathways nonspecifically, which may help to stabilize posture during interlimb coordination. These results may have implications for targeted neurorehabilitation after stroke.