Predicting the dynamics of network connectivity in the neocortex

Supplemental Information for: Loewenstein, Yanover and Rumpel, Predicting the dynamics of network connectivity in the neocortex, The Journal of Neuroscience, 09/2015, Volume 35, Issue 36, p.12535-12544 .

We have chronically imaged dendritic spines, the morphological correlates of excitatory synapses, on layer 5 pyramidal neurons in the auditory cortex of transgenic mice (GFP-M line, Feng et al., 2000, Neuron) through a small glass window using two-photon imaging. This approach allows monitoring a large fraction of the excitatory inputs on the apical tuft of individual pyramidal neurons. Interestingly, connectivity in the cortex is highly dynamic, while at the same time preserving stable elements.

Here we post the raw data analyzed in Loewenstein et al., 2011 and Loewenstein et al., 2015 as a table that contains the morphological variables of the spines, as well as their relative locations, as described below. The experimental procedures are described in details in Loewenstein et al., 2011 and Loewenstein et al., 2015.

Data

The data is structures as a 8,699 × 13 table, where each row corresponds to all properties of a single spine in a single imaging session. The data is available here in the comma-separated value (CSV) format.

• Columns 1-3 denote the identity of the spine:
• Column 1 denotes the index of the cell (there were 8 neurons in this dataset).
• Column 2 denotes the index of the dendrite in the cell.
• Column 3 denotes the index of the spine in the dendrite.
• Column 4 denotes the imaging session. There were six imaging sessions at an interval of four days.
• Columns 5-8 denote the morphological characteristics of the spines:
• Column 5 denotes the relative intensity of the spine — a measure of its volume (in arbitrary units).
• Columns 6 and 7 are measures of the shape of the spine: we performed a principal component analysis of the two-dimensional pixel map associated to the spine, where each pixel was weighted by its brightness. The values in columns 6 and 7 indicate the two eigenvalues λ₁ and λ₂, respectively, and the shape parameter used in the paper is

S := (λ₁-λ₂)/(λ₁+λ₂).

• Column 8 denotes the shortest distance of the center of mass of the spine to the center of the dendrite. This measure correlates with the length of a spine, however, can be an underestimation for particularly curved spines. The units are pixels (in the original images).
• Columns 9-13 denote the relative locations of the spines: This aspect of the data was not considered in Loewenstein et al., 2011 and Loewenstein et al., 2015. It is important to note that the location of the origin has changes between imaging sessions so that the absolute location of a spine is immaterial:
• Columns 9 and 10 denote the X-Y coordinates of the center of mass of the spine (in units of pixels).
• Columns 11 and 12 denote the X-Y coordinates of the point on the dendrite closest to the center of mass of the spine (in units of pixels).
• Column 13 denotes the location of the spine relative to the dendrite in the Z-direction: a 2D-analysis of spine morphology using two-photon microscopy can be confounded due to the lower vertical optical resolution, in particular when oblique or vertically extending dendritic spines are considered for analysis. For this reason, we focused our analysis on spines that are extending laterally from the dendrite, approximately parallel to the imaging plane. Considering those spines, column 13 denotes the difference in z-slices (corresponding to 0.5μm each) between the z-slice in which the brightness of the spine is highest and the z-slice in which the adjacent dendrite shows highest brightness.

Contact us

For additional information, please contact Yonatan Loewenstein (first name - dot - last name - at – mail.huji.ac.il) Uri Yanover (likewise) or Simon Rumpel ('si' - last name - at – 'uni-mainz.de').