An overview of Tactile Memory Tactile memory is part of sensory memory systems and it is the recollection of information acquired via touch. It is one of the primitive sensory codes that are used as interacting familiar objects. It is not only important to interact with familiar objects but it is also necessary to interact with novel objects with similar size.
Traces of tactile information is similar to iconic memory in terms of duration of the trace since it lasts for a short time and it is vulnerable to decay after almost two seconds (Gallace, &Spence, 2009). One of the earliest experimental studies on tactile memory was conducted by Bliss, Crane, Mansfield, and Townsend (1966). In this study, they investigated the characteristics of immediate recall for brief tactile stimuli applied to the hand.
The results obtained showed a haptic memory store remarkably similar to the visual memory store. Similar to tests of visual sensory memory, it was also found that haptic memory performance was significantly improved with the use of partial report procedures. In a recent study, Gallace and Spence (2009) also verified these findings. According to Bliss and colleagues, the difference between partial report and whole report is the result of a sensory form of memory for passively presented tactile stimuli with a high capacity and short duration.
Furthermore, Gilson and Baddeley (1969) argued that memory for stimuli applied to the skin is resilient for approximately ten seconds after removal of the stimulus, even when the individual is engaged in tasks that inhibit verbal rehearsal. After this delay, the memory trace becomes vulnerable to forgetting as it decays from the haptic memory store and begins to rely on a more central memory store. Although tactile memory representations can be thought as similar to visual representations in nature but there are significant differences between these two different memory systems in terms of processing and neural anatomy.
Easton, Srivinas, and Greene (1997) showed that there is an innate difference between visual and tactile memory representations. In their study, they presented their participants with an object either in visual or tactile forms. In the study, the participants viewed a sphere but they could not touch it. After that, the participants were given a similar shape but they could not see it. The results of the study indicated that the participants’ performance was worse as they were judging size differences in visual form compared to that in tactile form.
It is suggested that this is because the participants’ processing in visual form resulted in more variance in terms of object size due to influences, such as perspective and distance. Nero-anatomy of Tactile Memory Tactile memory is widely organized in the somatosensory cortex. The information signals received by body surface goes to the areas that are close together on the brain surface. Various areas of the parietal lobe are responsible for contributing to several aspects of tactile memory.
Memory for the features of a stimulus including its roughness, spatial density, and texture result in activation of the parietal operculum. On the other hand, features of a stimulus, such as size and shape are detected by touch receptors in the skin. These signals are maintained in the anterior part of the parietal lobe. In addition to features of the stimulus, memory for spatial information such as the location of stimulusactivates the right superior parietal lobule andtemporoparietal junction (Gallace, &Spence, 2008).
There are several studies investigating neural correlates of tactile memory. In one study, Harris, Harris, and Diamond (2001) assess the contribution of topographically organized neural areas to tactile working memory. In order to investigate neural correlates of tactile working memory, they presented their participants with vibrations in different frequencies. The participants were asked to compare the frequency of two vibrations. The vibrations were presented to either the same fingertip or to different fingertips. The retention interval between vibrations was at various lengths.
The results of the study indicated that participants performed well if their task was to compare vibrations delivered either the same finger or to corresponding fingers on opposite hands. However, their performance was lower when the vibrations were implemented on distant finger on either hand. These results demonstrate that tactile working memory mechanism organized in topographical framework. In another experiment, Harris and colleagues (2001) presented vibrations to the same fingertip but they added an interference vibration to the retention interval.
Participants’ task was again to compare the frequencies of two vibrations by ignoring interference vibration in retention interval. The results showed that the interpolated vibration impaired performance especially if it is delivered to the same finger as the comparison vibrations. The effect was smaller if the vibrations were delivered to more distant fingers. This finding verified the previous finding suggesting that tactile working memory is supported by topographically organized regions of somatosensory cortex. In a similar vein, Numminen et al. (2004) investigated tactile information in brief intervals.
Rather than longer interval durations, the study investigated the brain activation using functional magnetic resonance imaging (fMRI) in short inter-stimulus intervals. Participants were presented with triplets of pressure pulses to the second and fourth fingers of the left hand. After a brief interval, a second triplet was presented. In “compare” task, participants were asked to give feedback by finger lift if the triplets were same. The experiment also included a “control” task. The control task was same as the compare task but infrequent stimuli applied to the little finger to which the participants reacted.
There were different inter-stimulus intervals. The activation in the brain areas as a result of participants’ efforts in recognizing, maintaining and comparing tactile triplets in compare and control conditions showed the related brain areas. The results of the study demonstrated that activation in inferior parietal cortex, supplementary motor area, and right dorsolateral prefrontal cortex was enhanced during the compare task compared to the activations in the control task. It seems that the activation in the dorsolateral prefrontal cortex is the result of an attempt to memorize the stimulus sequence.
The activation in supplementary motor area and inferior parietal cortex is the result of an effort to analyze temporospatial tactile patterns and haptic exploration. The comparison task itself required a high effort and it resulted in an enhanced activation in the anterior cingulate gyrus. All in all, the results suggested that these areas revealed a task-specific activation. The process engaged in comparison task is common with normal processing of tactile stimuli and therefore, it can be said that dorsolateral prefrontal cortex, supplementary motor area and inferior parietal cortex are involved in processing tactile stimuli.
Another line of research focuses on the separation between and integration of spatiovisual and tactile memory systems. As previously mentioned, although tactile memory representations can be thought as similar to visual representations in nature but there are significant differences between these two different memory systems in terms of processing and neural anatomy. Saito et al (2003) used fMRI to assess the neural substrates for tactile-visual cross-modal matching. The task used in the experiment involved tactile – visual matching of two-dimensional shapes.
The participants performed four tasks. TT task required a tactile – tactile matching with no visual stimulus, TTv task required a tactile – tactile matching with visual stimuli, VVt task required a visual – visual matching with tactile stimuli, and TV task required a tactile – visual matching. It was expected that the neural substrates for tactile and visual shape matching were enhanced during tasks requiring matching of information coming from different sensory modalities, which are visual, and tactile modalities compared to the task requiring a matching within the same modality.
The results of the study showed that TT task activated the following areas: contralateral primary sensory motor area, post-central gyrus superior parietal lobules, anterior part of the intraparietal sulcus, thalamus, cerebellum, and supplementary motor area but there was no occipital involvement since there was no visual part in the task. Visual matching task, on the other hand, activated primary visual cortex, lingual and fusiform gyri. However, the tasks required cross-modal effort resulted in enhanced activation in the posterior intraparietal sulcus bilaterally.
This means that shape information coming from different modalities may be integrated in this region, the posterior intraparietal sulcus. These findings lead researchers to cross-modal integration between visual-spatial and tactile information. On the other hand, the study explained above used only 2D (two-dimensional) objects. However, using 3D (three-dimensional) stimuli may be more similar in real life situation. Using 3D stimuli may also enable to investigate a 3-way cross-modal integration that is the integration between visual – spatial-tactile modalities.
In the present study, it aimed to investigate the cross-modal integration of visual-spatial-tactile information by using 3D stimuli and its neural correlates by utilizing fMRI. Method Participants Ten healthy volunteers participated in this study. Eight of them were right handed and other two were left-handed. There was no history of neurological and psychiatric illness in any of the subjects. The ethical committee of Yeditepe University approved the protocol and all subjects gave their written informed consent for the study. Matching Task
For the tactile–tactile, visual–visual, or tactile–visual matching tasks, we used patterns of cube, rectangular parallelepiped, and cone. We used two cubes with different sizes – one’s edge was 3,5 cm and the other one’s edge was 5 cm. We also used a rectangular parallelepiped which had a long edge for 5 cm and short edge for 3,5 cm and last edge was 5 cm again. Lastly, we used 3 types of cone – one’s height was 5 cm, one’s height was 3,5 cm and last oblique cone’s height was 5 cm. Each pattern had one or three lobule on the surface, which can be noticeable either with tactilely or visually.
The subjects performed four different tasks: a tactile–tactile matching task with no visual input (TT), a tactile–tactile matching task with visual input (TTv), a visual–visual matching task with tactile input (VVt), and a tactile–visual matching task (TV). One task was performed in each fMRI sessions. Each task session was repeated twice, and thus, each subject completed eight sessions. The presentation order of the eight sessions was counterbalanced. Prior to the fMRI session, the subjects were trained for the tactile discrimination task.
For the TT task, subjects were asked to place their right hand in a supine position. Their left hand was placed on the button, which was connected to a microcomputer for recording their responses. The subjects closed their eyes throughout the session. During the task period a cube (Figure 1) was manually placed on the subject’s right palm. The subjects were required to explore the surface and edges of the cube with right hand’s fingers for 7 seconds.
Then, participants had another 7 seconds to explore the pair of the pattern before gave his/her response to the question “were they same? If the participant thinks they were same, then he/she used left index finger, if they weren’t same then he/she used left middle finger to push a button. After they responded, they were all allowed to drop the pattern. For the TTv task, the procedure was identical to the TT task except that additional visual 3D stimuli were presented simultaneously. The visual stimulation was projected using a television screen. The subjects through a mirror viewed the screen. It was confirmed that the subjects were not able to see their right hand.
During the task periods, the tactile and visual stimuli were presented simultaneously for 7 seconds, then, pairs of patterns were presented visually and tactilely for another 7 seconds, followed by a response. The subjects responded by pushing a button with the left index finger if the tactile pair-wise pattern were the same, or with the middle finger if the pattern were different, irrespective of the visual stimuli.