Ragni Flavio, Tucciarelli Raffaele, Andersson Patrik, Lingnau Angelika
Center for Mind/Brain Science (CIMeC), University of Trento, Rovereto, TN, Italy.
Center for Mind/Brain Science (CIMeC), University of Trento, Rovereto, TN, Italy; Department of Psychological Sciences, Birkbeck, University of London, London, UK.
Cortex. 2020 Jun;127:371-387. doi: 10.1016/j.cortex.2020.02.020. Epub 2020 Mar 18.
In the absence of input from the external world, humans are still able to generate vivid mental images. This cognitive process, known as visual mental imagery, involves a network of prefrontal, parietal, inferotemporal, and occipital regions. Using multivariate pattern analysis (MVPA), previous studies were able to distinguish between the different orientations of imagined gratings, but not between more complex imagined stimuli, such as common objects, in early visual cortex (V1). Here we asked whether letters, simple shapes, and objects can be decoded in early visual areas during visual mental imagery. In a delayed spatial judgment task, we asked participants to observe or imagine stimuli. To examine whether it is possible to discriminate between neural patterns during perception and visual mental imagery, we performed ROI-based and whole-brain searchlight-based MVPA. We were able to decode imagined stimuli in early visual (V1, V2), parietal (SPL, IPL, aIPS), inferotemporal (LOC) and prefrontal (PMd) areas. In a subset of these areas (i.e., V1, V2, LOC, SPL, IPL and aIPS), we also obtained significant cross-decoding across visual imagery and perception. Moreover, we observed a linear relationship between behavioral accuracy and the amplitude of the BOLD signal in parietal and inferotemporal cortices, but not in early visual cortex, in line with the view that these areas contribute to the ability to perform visual imagery. Together, our results suggest that in the absence of bottom-up visual inputs, patterns of functional activation in early visual cortex allow distinguishing between different imagined stimulus exemplars, most likely mediated by signals from parietal and inferotemporal areas.
在没有来自外部世界输入的情况下，人类仍然能够产生生动的心理意象。这种认知过程，即视觉心理意象，涉及前额叶、顶叶、颞下叶和枕叶区域组成的神经网络。使用多变量模式分析（MVPA），先前的研究能够区分想象光栅的不同方向，但无法在早期视觉皮层（V1）中区分更复杂的想象刺激，如常见物体。在这里，我们研究了在视觉心理意象过程中，字母、简单形状和物体是否能在早期视觉区域被解码。在一个延迟空间判断任务中，我们要求参与者观察或想象刺激。为了检验在感知和视觉心理意象过程中是否能够区分神经模式，我们进行了基于感兴趣区域（ROI）和全脑搜索光的MVPA。我们能够在早期视觉（V1、V2）、顶叶（顶上小叶、顶下小叶、前顶内沟）、颞下叶（外侧枕叶复合体）和前额叶（背侧运动前区）区域解码想象的刺激。在这些区域的一个子集中（即V1、V2、外侧枕叶复合体、顶上小叶、顶下小叶和前顶内沟），我们还在视觉意象和感知之间获得了显著的交叉解码。此外，我们观察到顶叶和颞下叶皮层中行为准确性与血氧水平依赖（BOLD）信号幅度之间存在线性关系，但在早期视觉皮层中没有，这与这些区域有助于视觉意象能力的观点一致。总之，我们的结果表明，在没有自下而上的视觉输入时，早期视觉皮层中的功能激活模式能够区分不同的想象刺激样本，这很可能由顶叶和颞下叶区域的信号介导。
