Neuroscientists often describe neural activity as a representation of something, or claim to
have found evidence for a neural representation, but there is considerable ambiguity about …

… (B) Here we illustrate how an example visual stimulus is reformatted as a time-dependent
vector. Each stimulus frame has two spatial coordinates and a total of NX pixels. First, the …

… The variables V 0 , V 0 a , and V 0 b represent the membrane potentials of soma, apical
dendrite and basal dendrite, respectively. The dendritic compartments are defined as weighted …

Background We present the genome sequence of the tammar wallaby, Macropus eugenii,
which is a member of the kangaroo family and the first representative of the iconic hopping …

… and deletion (or INDEL) variation of significance (Supplemental Experimental Procedures),
extreme structural rearrangement and copy number variation was apparent (Figure 2B). …

Highlights • Learning-to-learn makes learning better. • Over many tasks a learner can
develop learning strategies (learn to optimize). • And it can discover structure that is common …

… (B) The PCs are computed by finding the eigenvectors of the original covariance matrix (top).
When viewed in the reference frame of the PCs, all off-diagonal covariances become zero (…

… B through node perturbation then we converge to W. Validation that the method learns to
approximate W, for fixed W, is provided in the supplementary material. In practice, we update B …

When a neuron is driven beyond its threshold, it spikes. The fact that it does not communicate
its continuous membrane potential is usually seen as a computational liability. Here we …

… (B and C) (B) Images of L4 excitatory cells responding to sound with changes in GCaMP6f
fluorescence in the ACx. (C) Segmented ROIs representing individual excitatory neurons. …