The pairwise decoding had an average of 60.0% accuracy with specific pairs having an accuracy up to 70.3%. The data that Horikawa and his colleagues found, showed that the patterns of activation was more telling about the information of the dream imagery, than the activity level seen in any particular area. They also found that the higher visual cortex areas of the brain had a higher accuracy in predicting the reported synsets. In the multilabel decoding portion of the experiment similar results were found, where the higher visual cortex areas were more accurate in predicting the visual imagery reports of the participant. They also found that the fusiform face area showed better performance with human synsets, and the parahippocampal place area showed better performance with place synsets. These finding match the normal response to direct visual stimulus that has been demonstrated in previous neuropsychological studies. Further analyzation of the data, especially sets that had a high score for unreported synsets, shows that decoding systems may be recording reported and unreported visual imagery
The pairwise decoding had an average of 60.0% accuracy with specific pairs having an accuracy up to 70.3%. The data that Horikawa and his colleagues found, showed that the patterns of activation was more telling about the information of the dream imagery, than the activity level seen in any particular area. They also found that the higher visual cortex areas of the brain had a higher accuracy in predicting the reported synsets. In the multilabel decoding portion of the experiment similar results were found, where the higher visual cortex areas were more accurate in predicting the visual imagery reports of the participant. They also found that the fusiform face area showed better performance with human synsets, and the parahippocampal place area showed better performance with place synsets. These finding match the normal response to direct visual stimulus that has been demonstrated in previous neuropsychological studies. Further analyzation of the data, especially sets that had a high score for unreported synsets, shows that decoding systems may be recording reported and unreported visual imagery