Visually guided performance of simple tasks using simulated prosthetic vision
Visually guided performance of simple tasks using simulated prosthetic vision
复制标题
DOI：
10.1046/j.1525-1594.2003.07309.x
复制
发表时间：
2003-11-01
影响因子：
2.4
通讯作者：
Dagnelie, G
中科院分区：
文献类型：
--
作者：
Hayes, JS;Yin, VT;Dagnelie, G
光感受器细胞的丧失是失明的主要原因之一。几个小组正在探索用视网膜假体替代光感受器的功能。本研究的目的是模拟具有4x4、6x10和16x16电极阵列的视网膜假体的受者可能体验的视觉水平，并测试这种视觉的功能。一台PC摄像机捕捉到的图像被实时转换成点（“像素”）。测试对象佩戴的PC显示器和头戴式显示器显示像素化图像。为了评估正常视力个体在每个数组中的表现，我们设计了一组任务，包括斯隆字母E的四选择方向辨别、物体识别和辨别、切割任务、倾倒任务、符号识别和两个阅读任务。在字母E任务中，受试者在4x4、6x10和16x16阵列下的视敏度分别为20/1,810、20/1,330和20/420。大多数受试者能够在16x16阵列上阅读小至36点的字体，在我们的系统中对应的视觉灵敏度为20/600。测试对象部分克服了系统的视觉限制，通过扫描字母上的摄像机，允许视觉信息的空间和时间整合。在所有类别中，受试者在16x16阵列中表现最好，在4x4阵列中表现最差。然而，即使使用最低分辨率的阵列，一些受试者也能识别简单的物体和符号。
Loss of photoreceptor cells is one of the major causes of blindness. Several groups are exploring the functional replacement of photoreceptors by a retinal prosthesis. The goal of this study was to simulate the vision levels that recipients of retinal prostheses with 4x4, 6x10, and 16x16 electrode arrays may experience, and to test the functionality of this vision. A PC video camera captured images that were converted in real time into dots ("pixels"). The PC monitor and a head-mounted display worn by test subjects displayed the pixelized images. To assess performance of normally sighted individuals with each array, we designed a set of tasks including: four-choice orientation discrimination of a Sloan letter E, object recognition and discrimination, a cutting task, a pouring task, symbol recognition, and two reading tasks. In the letter E task, subjects were found to have visual acuities of 20/1,810, 20/1,330, and 20/420 with the 4x4, 6x10, and 16x16 arrays, respectively. Most subjects were able to read fonts as small as 36 point with the 16x16 array, corresponding with a visual acuity of 20/600 in our system. The test subjects partially overcame the visual limitation of the system by scanning the video camera over the letters allowing spatial and temporal integration of visual information. In all categories, subjects performed best with the 16x16 array and least well with the 4x4 array. Even with the lowest resolution array, however, some subjects could recognize simple objects and symbols.