The analysis portrayed in Figure 6 was made because it was conceivable that the well-above chance probability (reported in Figure 3) was driven by transfer of visual information, from seeing the hand making the motoric response, to the systems deciding which symbolic response to make. H. ChristiansenJens ChristensenJeppe GrünbaumThor KyllingsbækSøren 2014 <p>For such a transfer to occur, though, an association must be formed between visual information, from seeing the hand making the motoric response and a symbolic response. <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0094744#pone-0094744-g006" target="_blank">Figure 6</a> shows that, during the double response trials, no associations are formed between visual information, from seeing the hand making a motoric response and the corresponding symbolic response. If such associations were formed then the associations should become stronger and more precise over time as part of a learning process. This means that the well-above chance probability of making the same symbolic and motoric response should be increasing over time as well as one moves from block 1–4 during part 5 of the experimental procedure where the double responses are made. As can be readily observed, from visual inspection of the simple linear regression fitted to the data (red line), the difference between the observed probability (P(obs)) and the predicted probability (P(pred)) is not increasing as one moves from block 1 to block 4. The data-points were created by reanalyzing the data reported in <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0094744#pone-0094744-g003" target="_blank">Figure 3</a> (for Group 1). Only data from session 1 is part of the analysis (160 trials pr. participant). The 160 trials were presented in four blocks. Within blocks all combinations of length and orientation were shown. For length, each of the data-point consists of 10 trials (160 trials/4 blocks/4 lengths) and for orientation each data-point consists of 8 trials (160 trials/4 blocks/5 orientations). For Group 1 the slope of the red line was negative for both length (−0.025) and orientation (−0.0214). For Group 1, a one-way ANOVA found no significant effect of blocks for either length or orientation (Length: F(3,108) = 1.44, p = 0.236; Orientation: F(3,136) = 1.93, p = 0.127). For Group 2 the slope of the red line was negative for length (−0.057) but slightly positive for orientation (0.0089). A one-way ANOVA did find a significant effect of blocks on the negative slope for length (F(3,108) = 2.89, p = 0.0387) but no significant effect of blocks on the slightly positive slope for orientation (F(3,136) = 2.32, p = 0.0777). Thus, feedback, from the hand making the motoric response, was not driving the well-above chance occurrence of identical motoric and symbolic responses reported in <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0094744#pone-0094744-g003" target="_blank">Figure 3</a>.</p>