Abstract:
Mouse whisker-related primary somatosensory cortex (also known as barrel cortex, BCx) is
required to form an association between a behaviorally relevant tactile stimulus and its
consequences, only if the first conditioned stimulus CS (here a single whisker deflection),
and the latter unconditioned stimulus US (here a corneal air puff) are separated by a ‘trace’
(brief memory period). I investigated whether tactile trace eye blink conditioning (TTEBC)
has a correlate in BCx activity and whether such BCx activity in the two periods, CS and trace
are required for learning.
I trained three head-fixed mice on TTEBC to assess learning related functional plasticity of BCx
by recording LFPs and multi-unit (MU) spiking from 4-shank laminar silicone probes (8
electrodes per shank, inter-shank distance 200μm) spanning the depths of the principal barrel
column and its neighbors. Current source density analysis (CSD) showed the known short
latency sink (~8ms) in L4 and L5/6 during CS presentation, followed by a weaker current sink
during ongoing tactile stimulation, spanning across the column. At the same depth, a novel
current source was discovered during the trace period. The latter two currents were
consistently attenuated during TTEBC acquisition. Onset MU spike response to the CS (at a
latency of <15ms) was stable in most units, while steady state CS-response (50-250ms)
typically decreased below the pre-learning level. Spiking during the trace period also
depressed during learning. These plastic changes were observed in neighboring shanks at a
horizontal distance of up to 400μm. These findings show that BCx is functionally involved in
TTEBC acquisition. Matching the lateral spread of the neuronal signal into the neighboring
column, I found mice to generalize the CS-US association only to adjacent, but not to near and
far whiskers.
I next asked whether the involvement of BCx during the trace period has any causal role in
TTEBC. I employed the well-established VGAT-ChR2 mouse line that, due to expression of
channelrhodopsin-2 in inhibitory neurons (Zhao et al., 2011), blocks virtually all spikes in a
column with high temporal precision, using blue light. I found that BCx functionality was
required during CS presentation. However, mice learned normally when blocking BCx during
the trace period. After learning, BCx activity during CS & trace was entirely dispensable for
task performance.
In summary, I demonstrate that the barrel column is involved in acquiring the TTEBC
association. Nevertheless, the plasticity of the neuronal response in the trace period is a non-
causal reflection of learning, and after learning, in the early phase of retention BCx is not
needed for task performance. Future research need to establish if BCx assumes a more critical
role in late consolidation. Further, the nature and projection of the signals measured during
the learning have to be explored on the microscopic network and cellular level.