So RNA-Seq is a snapshot measurement of RNA fragments, in this case RNA gene expression, at any given time from a sample. The small black lines underneath the gene annotation corresponds to these RNA fragments and aligned to the exon sequence. For example, there is 2 black lines between exon 1 and 2 in the wildtype brain sample, this means after alignment, sections of this RNA is composed of bases from both exon 1 and exon 2. Whereas the 2 black lines under exon 2 alone means after alignment, these bases are only from exon 2.
Now you know this information, you can analyse the results. The key thing to note is that in the brain samples (both mutant and wild type), there are no black lines underneath the gene connected to exon 4 which is where the mutation is. From this information, you can safely deduce that the particular isoform of this protein produced in the brain does not require a functional exon 4, as exon 4 is spliced out. You can also observe that the RNA-Seq from the mutant brain tissue and wild-type brain tissue are exactly the same, therefore the protein is completely unaffected in the brain.
Next, you can compare the wild-type eye to the mutant eye. Firstly, in the wild-type eye, there's 6 RNA fragments which include at least part of exon 4. But when you compare this to the mutant eye, only 4 RNA fragments contain sections of exon 4. You can also tell these 4 RNA sequences are in different positions to the 6 from the wild-type. So clearly this mutation is affecting the gene transcripts (both quantity and positions) and only in the eye, so it wouldn't be a stretch to say the molecular mechanism caused by faulty protein production in the eyes and not the brain.