Go read. I have nothing else to say other than: Bravo!!
UPDATE:
I have a few minutes so I would like to highlight this paragraph written by editor in chief Dr. John Maunsell:
Another troubling problem associated with supplemental material is that it encourages excessive demands from reviewers. Increasingly, reviewers insist that authors add further analyses or experiments “in the supplemental material.” These additions are invariably subordinate or tangential, but they represent real work for authors and they delay publication. Such requests can be an unjustified burden on authors. In principle, editors can overrule these requests, but this represents additional work for the editors, who may fail to adequately referee this aspect of the review.
In my opinion this is absolutely correct and gets right to the heart of what I think is wrong with science today. Our papers are ultimately about ideas and the experiments that either support or reject those ideas. The constant you just need one more piece of supporting evidence for everything mindset of many reviewers (I include myself in falling into this trap) is not useful to the process. The ideas in a given paper may or may not stand the test of time and that one more piece of supporting evidence is unlikely to have any influence on what that test of time will determine. The point is to get potentially influential ideas out there and to get them out earlier rather than later. Post publication experimental scrutiny is and will always be how the test of time determines the validity of new scientific concepts.
UPDATE II:
DM agrees
JUNIORPROF 
HOLY CRAP!! That is the best news I’ve heard in a long time. Supplemental data is getting ridiculous. I hope more journals follow suit. Hooray for J. Neuroscience!!!!
This is a fantastic step to excise the GlamourMag cancer! Go J Neuro!!!!!!!
I’m not in this field but this strikes me as dated and taking a hands off approach to the (potentially negative) impact of a given study. My read from the editors would be a resounding antagonistic view of many of these high data papers: it’s crap, don’t believe it, we can’t say no the week we read it because of practicalities, we published it, and then it turns out the work is shite a few months later. Buyers remorse or what?
Aren’t we in the age of multi-parameter, 3D googledy-goggly-omics whose documentation at an established source keeps the early adopters at least somewhat in check? Are we to take lists of data (or the top 20 of lists of lists) published in Table 3 as truth without seeing at least parts of the giggety-giggety bytes of data behind those conclusions?
I’ve recently switched from a reductionist approach to a wishy-washy-omics/living cell picture and can easily see how “massaging” data lends to …well…lets just say “new findings” that could be construed as correct given parameters that are …. well … not generally valid in a living organism.
This is great news. A colleague told me recently that after a year of back-and-forth at Nature Neuroscience his paper was accepted. During the several review stages more and more supplementary info had to be added (but changed little in the actual paper). The paper, now accepted, had 27 pages of supplemental info. NN said “you have to cut it to 5 pages”. So what was the point in wasting a year doing experiments at the whim of reviewers if nobody is going to be able to look at the results? Best thing is to get rid of it altogether. To answer anonymous, it’s not really a backwards step. Neuroscientists accept that they can’t download other people’s traces and re-analyse them. And there ain’t many systems level papers published in this Journal!
I hope other journals (specifically those in my field) take the cue.
I don’t think this is a sensible idea. In my field (chemistry/chemical biology), some data/figure/table are not of interest to the “general readership” but they are essential for people in the field especially if they want to repeat a certain experiment. Things such as reaction conditions/detailed characterisation of compounds/sequence alignment etc. Yes it maybe possible to find some of these information from databases etc but as a researcher it saves me time looking online for these essential information – in fact, in some papers, the supplemental info is actually more useful than the main text. The main text is interesting in presenting an idea/story but the supplemental info is essential for following up and repeating the experiments
I agree, Anonymous2 – just yesterday I dug into the supplementary info on a paper because I wanted more detail on the control group conditions. Surely the supplementary information could be provided under less rigorous conditions to make it more manageable. Now I’ll have to start harassing researchers by email to get the information I need, which is less than ideal when you’re a student. Most people are generous with their time, but there’s a limit to how many undergraduate questions they have time for.