How do function and projection targets constrain synaptic connectivity?
COMPUTATION
Valverde (J Neurocytology 1976)
Multiple whole-cell recordings to assay connections between cell pairs
• Information about the presence and strength of connections
• But false negatives due to slicing artefacts
Song, Sjöström, Reigl, Nelson & Chklovskii (PLoS Biology 2005)
Excitatory connectivity in the neocortex is neither uniform nor random
Overrepresentation of
bidirectional motifs
Connection strengths are variable
and their distribution has a long tail
L5 pyramidal cells, Song et al. (2005)
L5/6 pyramidal cells, Deuchars, West, Thomson (1994)
Vibrissa cortex, all excitatory connections
Visual cortex, L2/3 à L2/3 pyramid
Lefort et al 2009
Holmgren et al 2003
Vibrissa cortex, L4 stellate à L2/3 pyramid
Vibrissa cortex, L5 à L5 pyramid
Feldmeyeret al 2002
Markram et al 1997
Are neurons connected by few strong synapses in a sea of weak synapses?
Layer 2/3 pyramidal cell network
Strong
Weak
Cossell, Iacaruso, Muir, Houlton, Sader, Ko, Hofer & Mrsic-Flogel (Nature 2015)
Acuity (cycle/°)
Human 46
Mouse 0.5
Mapping
with two-photon calcium imaging and reverse correlation
Example cell
50 µm
Stimuli
Linear
Calcium response
20 o
*
…
…
…
=
Quiescence
20%
dF/F
Spike
Cossell, Iacaruso, Muir, Houlton, Sader, Ko, Hofer & Mrsic-Flogel (Nature 2015)
Mutiple patch clamp
slice
Layer 2/3
300 μm x 300 μm x 56 μm
Cossell, Iacaruso, Muir, Houlton, Sader, Ko, Hofer & Mrsic-Flogel (Nature 2015)
Relating synaptic connections to receptive fields
20 o
Cell 5
Cell 4
Cell 2
Cell 3
Cell 1
20µm
Exp : 20130522, Motif 2
Cossell, Iacaruso, Muir, Houlton, Sader, Ko, Hofer & Mrsic-Flogel (Nature 2015)
Spatial correlation of the receptive fields as a measure of cell-to-cell similarity
x 10 4
7
6
# occurrences
5
4
3
-0.56
0.51
2
1
0
-0.8 -0.6 -0.4 -0.2
0
0.2
Spatial correlation
0.4
0.6
Cossell, Iacaruso, Muir, Houlton, Sader, Ko, Hofer & Mrsic-Flogel (Nature 2015)
0.6
0.5
0.4
0.3
4
4/6
EPSP amplitude (mV)
Connection probability
0.7
10/32
1/4
19/87
0.2
0.1
0
-0.8
7/63
2/27
17/128
3.5
0.8
6
5
3
2.5
4
2
3
1.5
2
1
1
0.5
15/173
-0.4
0
0.4
Spatial correlation
Connection type:
Bidirectional
Unidirectional
0
-0.8
-0.4
0
0.4
Spatial correlation
0
0.8
Density function of spatial correlations
Similarity of receptive fields predicts probability, strength,
and reciprocity of synaptic connections
• Spatial correlation is a strong predictor of connectivity
• Cell pairs with positive correlations are more likely to connect with strong connections
• Reciprocal connections are stronger and exist between cell pairs with similar receptive fields
Cossell, Iacaruso, Muir, Houlton, Sader, Ko, Hofer & Mrsic-Flogel (Nature 2015)
At least in primary visual cortex, the evidence implies that
neurons are connected by a few strong synapses in a sea of weak synapses
Strong
Weak
Cossell, Iacaruso, Muir, Houlton, Sader, Ko, Hofer & Mrsic-Flogel (Nature 2015)