- Email: [email protected]
Human cortical response to incoherent motion on a background of coherent motion
Neuroscience Letters 347 (2003) 41–44 www.elsevier.com/locate/neulet
Human cortical response to incoherent motion on a background of coherent motion Khanh Lam, Yoshiki Kaneoke*, Ryusuke Kakigi Department of Integrative Physiology, National Institute for Physiological Sciences, Myodaiji, Okazaki 444-8585, Japan Received 28 April 2003; received in revised form 19 May 2003; accepted 19 May 2003
Abstract To investigate whether humans achieve a high sensitivity to coherent motion by excluding the response to incoherent motion, we measured the magnetoencephalographic response to the motion of randomly located dots one half of which moved coherently while the other half moved incoherently. The response was related to the faster motion of either coherent or incoherent motion though the observers saw both. All the estimated response sources were within the extrastriate area. The results indicate that incoherent motion is represented in the neural activity of the human extrastriate area even when the coherent motion is perceived at the same time. The fact that the neural activity for the slower coherent motion is not represented in the magnetic response suggests the existence of interaction between the neural activities for the two motions. q 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Human visual system; Extrastriate area; MT/V5; Magnetoencephalography; Random dot kinematogram; Coherent and incoherent motion
The coherent motion of dots among other randomly moving dots can be extracted by humans and monkeys even though only a few percent of the dots move coherently [1,15]. The extraction of the coherent motion from such a small amount of local motion is possible only when the direction information is integrated in a large spatial scale [19]. Incoherent motion does not have a global motion direction because all the dots move in random directions. Thus, it has been thought that the neurons in the extrastriate area do not respond to incoherent motion as much as to coherent motion. Recent physiological studies, however, have shown that MT/V5 of both monkey and human is activated by incoherent motion [2,3,6,11– 13]. These results apparently contradict the idea that MT/V5 is exclusively involved in the extraction of coherent motion in a background of random local motion such as that of incoherent motion. These human studies compared the responses to incoherent motion with those to coherent motion. Thus, the response of the extrastriate area to incoherent motion may occur only in the absence of coherent motion in the same visual scene. Six right-handed colleagues (four men and two women, aged 27 –42 years) participated in this study. Both coherent * Corresponding author. Tel.: þ 81-564-55-7766; fax: þ81-564-52-7913. E-mail address: [email protected] (Y. Kaneoke).
and incoherent motion were created using a random dot kinematogram (RDK) as in our previous study [10,12]. The stimuli were presented on a screen in a magnetically shielded room from outside through a small window using a Liquid Crystal Display projector (LP-9200; SANYO, Tokyo, Japan). The stimuli subtended a 13.5 £ 13.58 visual angle (1.1 cd/m2) with a dot size of 0.13 £ 0.138 and a density of 10% at a viewing distance of 2 m. Six different types of motion (coherent, incoherent, or both) were presented 2 –3 s after presenting the stationary dots. A 2 s motion presentation was followed by dots that stayed at the same locations for 2– 3 s and then started to move again. When both types of motion were presented (mixed motion), 50% of the dots moved coherently (C) at a speed of either 1.3 or 10 deg/s and the other 50% of the dots moved incoherently (I) at a different speed. These two speeds were chosen because the response latency for motion at 10 deg/s was about 50 ms shorter than that for motion at 1.3 deg/s according to our previous study [12]. Thus, we should be able to evaluate the response properties for the mixed motion by comparing them with the response properties for each separate type of motion. If the response to the mixed motion is a simple summation of the responses to the coherent and incoherent motion, the response should have two peaks with a 50 ms difference. If the coherent motion is
0304-3940/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0304-3940(03)00617-7
42
K. Lam et al. / Neuroscience Letters 347 (2003) 41–44
mainly responsible for the magnetic response, we should find that the first peak of the response to the mixed motion has a latency the same as that for the response to coherent motion of the same speed as that of the mixed motion. In this text, we refer to the motion types as C(1.3) þ I(10), which stands for the mixed motion of coherent motion at 1.3 deg/s and incoherent motion at 10 deg/s. When either coherent or incoherent motion was presented, 50% of the dots were used for the motion and the remaining dots were stationary (S). Such a pattern of motion is indicated as C(1.3) þ S, I(10) þ S, and so on. We used a 37-channel neuromagnetometer (Magnes; BTi, San Diego, CA) to record the magnetic responses from the right occipito-parieto-temporal region of each subject’s brain as in our previous study [12]. When the two types of motion were presented simultaneously, all the subjects had vivid perception of the coherent and incoherent motion as if they were on different planes. The magnetic responses to the mixed motion, however, were quite similar to those to the higher speed motion (10 deg/s). Fig. 1 shows the waveforms of the magnetic responses to the six types of motion with the time courses of the RMS values. The responses consisted of two to three components and the first component was always the largest. The peak latency of the response to the slower motion was about 50 ms longer than that for the faster motion. The waveforms and the time courses of the RMS values of the responses to the mixed motion were apparently the same as those for the higher speed motion irrespective of the motion type. Fig. 2 shows the peak latencies and the amplitudes of the responses to all the six motion stimuli for each subject. Two-way ANOVA (six subjects £ six stimulus conditions) revealed a significant effect (P , 0:001) of the stimulus condition on both latency and amplitude change (d:f: ¼ 5, F ¼ 29:5 and d:f: ¼ 5, F ¼ 6:6, respectively). The peak latency of the response to the slow speed motion was
Fig. 1. Averaged waveforms of the magnetic response to the onset of the motion and the RMS values from one subject (S1). The motion started at time zero. Waveforms from the 37 channels were overplayed at the mean baseline before the motion. A schematic illustration of the motion is presented at the left side of each response. Shaded dots are only for illustration. The arrowhead on RMS indicates the peak latency of the first component of the magnetic response. Note that the responses to the mixed motion (C þ I) are quite similar to the responses to the higher speed motion. That is, the response to C(10) þ I(1.3) is similar to that for C(10) þ S in terms of the response latency and the shape of the waveform. There is no trace of the response to the slower motion I(1.3) in the response to C(10) þ I(1.3).
Fig. 2. Changes in the peak latency and the amplitude of the first response component for each subject. All the data for both onset and offset responses are shown. The significance of the effect of the motion type revealed by two-factor ANOVA (subject £ stimulus) on the latency and the amplitude data is P , 0:001 (d:f: ¼ 5, F ¼ 29:5 and d:f: ¼ 5, F ¼ 6:6, respectively). Pair wise comparison by post-hoc test with Bonferroni adjustment revealed significant differences between some pairs as shown by the symbols (*P , 0:05; **P , 0:005; ***P , 0:001). The response latencies for the slower motions were significantly longer than those for the faster motions. The latencies for the mixed motions (C þ I) were the same as those for the higher motions irrespective of the motion types.
significantly longer (P , 0:001 by post-hoc test with Bonferroni’s method) than those to the higher speed motion and mixed motion when the three types of motion, the mixed motion and the two separately presented motions were compared. There was no difference in the response latency and amplitude between the mixed and the faster motion. When the response properties were compared between the two types of motion at the same speed, the latencies for the incoherent motion were significantly longer than those for the coherent motion (P , 0:05 by post-hoc test for the speed of 1.3). The locations for four subjects (S1 – S4) were around the temporo-parieto-occipital area including human MT þ and corresponded to the temporooccipital type of our previous study [4] (see Fig. 3). The other two subjects’ locations were more anterior and were around the posterior end of the superior temporal sulcus, which corresponded to the parietal type. The 95% confidence distance for each subject was between 1.1 and 2.4 cm. There was no significant difference (P . 0:05 by
K. Lam et al. / Neuroscience Letters 347 (2003) 41–44
Fig. 3. The estimated response sources for each stimulus condition. All the data successfully estimated with the single equivalent current dipole model are presented. Some subjects’ data (S1–S4) show that the locations for the same motion type tend to gather; there was no significant difference in the locations between the faster motions and the mixed motions. The origin of the coordinate system is the midpoint between the left and right preauricular (PA) points. The positive X-axis extends from the origin through the nasion. The positive Z-axis extends from the origin through the top of the head, such that it is perpendicular to the plane formed by the nasion and both PA points. The positive Y-axis extends from the origin through the left side of the head such that it is perpendicular to the X- and Zaxes.
discriminant analysis) in the estimated locations when the data for the faster motions were compared with those for the mixed motions. The fact that our incoherent motion stimulus evoked a response similar to that for the coherent motion indicates that the effect of the motion opponency [8] on the incoherent motion was negligible. This suggests that the motion opponency occurs only when two motions with opposite directions exist in a very local area. Our results apparently contradict those of a previous functional MRI study [16] which showed that the human MT þ activation was related to the motion coherency of random dots. The discrepancy may be explained by differences in the tasks. Their subjects had to judge the direction of global motion. In contrast, our subjects perceived the onset of any kind of motion. A recent psychophysical study [9] showed that the neural population for the detection of motion is different from that for direction discrimination. The responses we measured in the present study cannot simply be due to the local motions of the dots, because the response latency for the incoherent motion was significantly longer than that for the coherent motion when the values were compared at the same speed (1.3 deg/s) (see Fig. 2). In our previous study [12], we found that the response to incoherent motion had a significantly smaller amplitude than the response to coherent motion. The response latency for the incoherent motion tended to be longer than that for the coherent motion though the difference was not
43
significant. This fact indicates that the motion type affected the response. The difference in the two types of motion is detected by the spatiotemporal integration of the dots’ local motions because there was no difference in the other visual information, such as the contrast, the dots’ size and density, temporal frequency, and local motion speed. Thus, the response must have occurred after the spatiotemporal integration of the stimuli as we discussed in our previous studies [11,12]. It is not plausible that the spatiotemporal integration occurs only in the detection of coherent motion and that incoherent motion is simply perceived as numerous local motions. First, the response latency for the incoherent motion was longer than that for the coherent motion. Because the neural process for the spatiotemporal integration is considered to occur after the local motion detection [19], the response latency for the coherent motion would be longer than that for the incoherent motion only if local motion detection is necessary to perceive incoherent motion. Second, the estimated source of the response to the incoherent motion was in the extrastriate area as was that of the coherent motion. The neurons in the extrastriate area including MT þ have larger classical receptive fields than those in the primary visual cortex [18]. This is why the major role of MT and the other extrastriate areas is considered to be that of integrating local motions to detect global motions [7,17]. Thus, the main response to the incoherent motion should originate from the primary visual cortex if there is no spatiotemporal integration process in the detection of incoherent motion. When the coherent motion was presented on the background of slower incoherent motion, the magnetic response to this stimulus, C(10) þ I(1.3), was not different from that to the coherent motion, C(10) þ S. This result was expected based on the idea that the global motion (coherent motion) is extracted from the background noise (incoherent motion) and is represented in the neural activity of the extrastriate area. If this is the case, we should have seen the same result in the stimuli of C(1.3) þ I(10). Even though the local speed of the coherent motion was slower than that of the background incoherent motion, the global motion energy for C(1.3) is larger than that for I(10) because the incoherent motion does not have a global direction. The response to C(1.3) þ I(10), however, was not different from the response to I(10) þ S. This could not have been because the slow coherent motion was not detected. All the subjects perceived the coherent motion in the stimulus of C(1.3) þ I(10) vividly because the 50% coherency was much higher than the detection threshold (usually a few percent) [1,5,14]. Furthermore, the results cannot be simply explained by the idea that the faster motions would be responsible for the magnetic response irrespective of the motion type. Although the local motion speed for I(10) was faster than that of C(1.3), the response is considered to occur after the spatiotemporal summation of the stimuli as discussed above. Again, the global motion energy for
44
K. Lam et al. / Neuroscience Letters 347 (2003) 41–44
C(1.3) is larger than that for I(10). Thus, the result indicates that coherent motion is not exclusively represented in the neural activity of the human extrastriate area and that the neural activity for incoherent motion also exists in that area. The results that the responses to the mixed motions were similar to those for the faster motions irrespective of the motion type show that the responses to the slower motions in the mixed motions were suppressed by the faster motions of the different types. Thus, the neural activity representing the coherent and incoherent motions may interact in the extrastriate area, though the two activities may occur in the different neural populations as suggested in our previous study [12].
Acknowledgements We thank Mr O. Nagata and Y. Takeshima for technical support.
References [1] C.L. Baker Jr., R.F. Hess, J. Zihl, Residual motion perception in a “motion-blind” patient, assessed with limited-lifetime random dot stimuli, J. Neurosci. 11 (1991) 454–461. [2] O.J. Braddick, J.M.D. O’Brien, J. Wattam Bell, J. Atkinson, T. Hartley, Brain areas sensitive to coherent visual motion, Perception 30 (2001) 61–72. [3] K.H. Britten, M.N. Shadlen, W.T. Newsome, J.A. Movshon, Responses of neurons in macaque MT to stochastic motion signals, Vis. Neurosci. 10 (1993) 1157–1169. [4] M. Bundo, Y. Kaneoke, S. Inao, J. Yoshida, A. Nakamura, R. Kakigi, Human visual motion areas determined individually by magnetoencephalography and 3D magnetic resonance imaging, Hum. Brain Mapp. 11 (2000) 33–45. [5] S. Celebrini, W.T. Newsome, Neuronal and psychophysical sensitivity to motion signals in extrastriate area MST of the macaque monkey, J. Neurosci. 14 (1994) 4109–4124.
[6] K. Cheng, H. Fujita, I. Kanno, S. Miura, K. Tanaka, Human cortical regions activated by wide-field visual motion: an H2(15)O PET study, J. Neurophysiol. 74 (1995) 413–427. [7] M.S. Graziano, R.A. Andersen, R.J. Snowden, Tuning of MST neurons to spiral motions, J. Neurosci. 14 (1994) 54–67. [8] D.J. Heeger, G.M. Boynton, J.B. Demb, E. Seidemann, W.T. Newsome, Motion opponency in visual cortex, J. Neurosci. 19 (1999) 7162– 7174. [9] K. Hol, S. Treue, Different populations of neurons contribute to the detection and discrimination of visual motion, Vis. Res. 41 (2001) 685 –689. [10] Y. Kaneoke, M. Bundou, S. Koyama, H. Suzuki, R. Kakigi, Human cortical area responding to stimuli in apparent motion, NeuroReport 8 (1997) 677 –682. [11] K. Lam, Y. Kaneoke, A. Gunji, H. Yamasaki, E. Matsumoto, T. Naito, R. Kakigi, Magnetic response of human extrastriate cortex in the detection of coherent and incoherent motion, Neuroscience 97 (2000) 1–10. [12] K. Maruyama, Y. Kaneoke, K. Watanabe, R. Kakigi, Human cortical responses to coherent and incoherent motion as measured by magnetoencephalography, Neurosci. Res. 44 (2002) 193–203. [13] D.J. McKeefry, J.D.G. Watson, R.S. Frackowiak, K. Fong, S. Zeki, The activity in human areas V1/V2, V3, and V5 during the perception of coherent and incoherent motion, Neuroimage 5 (1997) 1–12. [14] W.T. Newsome, K.H. Britten, J.A. Movshon, Neuronal correlates of a perceptual decision, Nature 341 (1989) 52–54. [15] W.T. Newsome, E.B. Pare, A selective impairment of motion perception following lesions of the middle temporal visual area (MT), J. Neurosci. 8 (1988) 2201– 2211. [16] G. Rees, K. Friston, C. Koch, A direct quantitative relationship between the functional properties of human and macaque V5, Nat. Neurosci. 3 (2000) 716 –723. [17] H. Sakata, H. Shibutani, Y. Ito, K. Tsurugai, Parietal cortical neurons responding to rotary movement of visual stimulus in space, Exp. Brain Res. 61 (1986) 658– 663. [18] D.C. Van Essen, J.H. Maunsell, J.L. Bixby, The middle temporal visual area in the macaque: myeloarchitecture, connections, functional properties and topographic organization, J. Comp. Neurol. 199 (1981) 293 –326. [19] D.W. Williams, R. Sekuler, Coherent global motion percepts from stochastic local motions, Vis. Res. 24 (1984) 55–62.
Human cortical response to incoherent motion on a background of coherent motion Khanh Lam, Yoshiki Kaneoke*, Ryusuke Kakigi Department of Integrative Physiology, National Institute for Physiological Sciences, Myodaiji, Okazaki 444-8585, Japan Received 28 April 2003; received in revised form 19 May 2003; accepted 19 May 2003
Abstract To investigate whether humans achieve a high sensitivity to coherent motion by excluding the response to incoherent motion, we measured the magnetoencephalographic response to the motion of randomly located dots one half of which moved coherently while the other half moved incoherently. The response was related to the faster motion of either coherent or incoherent motion though the observers saw both. All the estimated response sources were within the extrastriate area. The results indicate that incoherent motion is represented in the neural activity of the human extrastriate area even when the coherent motion is perceived at the same time. The fact that the neural activity for the slower coherent motion is not represented in the magnetic response suggests the existence of interaction between the neural activities for the two motions. q 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Human visual system; Extrastriate area; MT/V5; Magnetoencephalography; Random dot kinematogram; Coherent and incoherent motion
The coherent motion of dots among other randomly moving dots can be extracted by humans and monkeys even though only a few percent of the dots move coherently [1,15]. The extraction of the coherent motion from such a small amount of local motion is possible only when the direction information is integrated in a large spatial scale [19]. Incoherent motion does not have a global motion direction because all the dots move in random directions. Thus, it has been thought that the neurons in the extrastriate area do not respond to incoherent motion as much as to coherent motion. Recent physiological studies, however, have shown that MT/V5 of both monkey and human is activated by incoherent motion [2,3,6,11– 13]. These results apparently contradict the idea that MT/V5 is exclusively involved in the extraction of coherent motion in a background of random local motion such as that of incoherent motion. These human studies compared the responses to incoherent motion with those to coherent motion. Thus, the response of the extrastriate area to incoherent motion may occur only in the absence of coherent motion in the same visual scene. Six right-handed colleagues (four men and two women, aged 27 –42 years) participated in this study. Both coherent * Corresponding author. Tel.: þ 81-564-55-7766; fax: þ81-564-52-7913. E-mail address: [email protected] (Y. Kaneoke).
and incoherent motion were created using a random dot kinematogram (RDK) as in our previous study [10,12]. The stimuli were presented on a screen in a magnetically shielded room from outside through a small window using a Liquid Crystal Display projector (LP-9200; SANYO, Tokyo, Japan). The stimuli subtended a 13.5 £ 13.58 visual angle (1.1 cd/m2) with a dot size of 0.13 £ 0.138 and a density of 10% at a viewing distance of 2 m. Six different types of motion (coherent, incoherent, or both) were presented 2 –3 s after presenting the stationary dots. A 2 s motion presentation was followed by dots that stayed at the same locations for 2– 3 s and then started to move again. When both types of motion were presented (mixed motion), 50% of the dots moved coherently (C) at a speed of either 1.3 or 10 deg/s and the other 50% of the dots moved incoherently (I) at a different speed. These two speeds were chosen because the response latency for motion at 10 deg/s was about 50 ms shorter than that for motion at 1.3 deg/s according to our previous study [12]. Thus, we should be able to evaluate the response properties for the mixed motion by comparing them with the response properties for each separate type of motion. If the response to the mixed motion is a simple summation of the responses to the coherent and incoherent motion, the response should have two peaks with a 50 ms difference. If the coherent motion is
0304-3940/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0304-3940(03)00617-7
42
K. Lam et al. / Neuroscience Letters 347 (2003) 41–44
mainly responsible for the magnetic response, we should find that the first peak of the response to the mixed motion has a latency the same as that for the response to coherent motion of the same speed as that of the mixed motion. In this text, we refer to the motion types as C(1.3) þ I(10), which stands for the mixed motion of coherent motion at 1.3 deg/s and incoherent motion at 10 deg/s. When either coherent or incoherent motion was presented, 50% of the dots were used for the motion and the remaining dots were stationary (S). Such a pattern of motion is indicated as C(1.3) þ S, I(10) þ S, and so on. We used a 37-channel neuromagnetometer (Magnes; BTi, San Diego, CA) to record the magnetic responses from the right occipito-parieto-temporal region of each subject’s brain as in our previous study [12]. When the two types of motion were presented simultaneously, all the subjects had vivid perception of the coherent and incoherent motion as if they were on different planes. The magnetic responses to the mixed motion, however, were quite similar to those to the higher speed motion (10 deg/s). Fig. 1 shows the waveforms of the magnetic responses to the six types of motion with the time courses of the RMS values. The responses consisted of two to three components and the first component was always the largest. The peak latency of the response to the slower motion was about 50 ms longer than that for the faster motion. The waveforms and the time courses of the RMS values of the responses to the mixed motion were apparently the same as those for the higher speed motion irrespective of the motion type. Fig. 2 shows the peak latencies and the amplitudes of the responses to all the six motion stimuli for each subject. Two-way ANOVA (six subjects £ six stimulus conditions) revealed a significant effect (P , 0:001) of the stimulus condition on both latency and amplitude change (d:f: ¼ 5, F ¼ 29:5 and d:f: ¼ 5, F ¼ 6:6, respectively). The peak latency of the response to the slow speed motion was
Fig. 1. Averaged waveforms of the magnetic response to the onset of the motion and the RMS values from one subject (S1). The motion started at time zero. Waveforms from the 37 channels were overplayed at the mean baseline before the motion. A schematic illustration of the motion is presented at the left side of each response. Shaded dots are only for illustration. The arrowhead on RMS indicates the peak latency of the first component of the magnetic response. Note that the responses to the mixed motion (C þ I) are quite similar to the responses to the higher speed motion. That is, the response to C(10) þ I(1.3) is similar to that for C(10) þ S in terms of the response latency and the shape of the waveform. There is no trace of the response to the slower motion I(1.3) in the response to C(10) þ I(1.3).
Fig. 2. Changes in the peak latency and the amplitude of the first response component for each subject. All the data for both onset and offset responses are shown. The significance of the effect of the motion type revealed by two-factor ANOVA (subject £ stimulus) on the latency and the amplitude data is P , 0:001 (d:f: ¼ 5, F ¼ 29:5 and d:f: ¼ 5, F ¼ 6:6, respectively). Pair wise comparison by post-hoc test with Bonferroni adjustment revealed significant differences between some pairs as shown by the symbols (*P , 0:05; **P , 0:005; ***P , 0:001). The response latencies for the slower motions were significantly longer than those for the faster motions. The latencies for the mixed motions (C þ I) were the same as those for the higher motions irrespective of the motion types.
significantly longer (P , 0:001 by post-hoc test with Bonferroni’s method) than those to the higher speed motion and mixed motion when the three types of motion, the mixed motion and the two separately presented motions were compared. There was no difference in the response latency and amplitude between the mixed and the faster motion. When the response properties were compared between the two types of motion at the same speed, the latencies for the incoherent motion were significantly longer than those for the coherent motion (P , 0:05 by post-hoc test for the speed of 1.3). The locations for four subjects (S1 – S4) were around the temporo-parieto-occipital area including human MT þ and corresponded to the temporooccipital type of our previous study [4] (see Fig. 3). The other two subjects’ locations were more anterior and were around the posterior end of the superior temporal sulcus, which corresponded to the parietal type. The 95% confidence distance for each subject was between 1.1 and 2.4 cm. There was no significant difference (P . 0:05 by
K. Lam et al. / Neuroscience Letters 347 (2003) 41–44
Fig. 3. The estimated response sources for each stimulus condition. All the data successfully estimated with the single equivalent current dipole model are presented. Some subjects’ data (S1–S4) show that the locations for the same motion type tend to gather; there was no significant difference in the locations between the faster motions and the mixed motions. The origin of the coordinate system is the midpoint between the left and right preauricular (PA) points. The positive X-axis extends from the origin through the nasion. The positive Z-axis extends from the origin through the top of the head, such that it is perpendicular to the plane formed by the nasion and both PA points. The positive Y-axis extends from the origin through the left side of the head such that it is perpendicular to the X- and Zaxes.
discriminant analysis) in the estimated locations when the data for the faster motions were compared with those for the mixed motions. The fact that our incoherent motion stimulus evoked a response similar to that for the coherent motion indicates that the effect of the motion opponency [8] on the incoherent motion was negligible. This suggests that the motion opponency occurs only when two motions with opposite directions exist in a very local area. Our results apparently contradict those of a previous functional MRI study [16] which showed that the human MT þ activation was related to the motion coherency of random dots. The discrepancy may be explained by differences in the tasks. Their subjects had to judge the direction of global motion. In contrast, our subjects perceived the onset of any kind of motion. A recent psychophysical study [9] showed that the neural population for the detection of motion is different from that for direction discrimination. The responses we measured in the present study cannot simply be due to the local motions of the dots, because the response latency for the incoherent motion was significantly longer than that for the coherent motion when the values were compared at the same speed (1.3 deg/s) (see Fig. 2). In our previous study [12], we found that the response to incoherent motion had a significantly smaller amplitude than the response to coherent motion. The response latency for the incoherent motion tended to be longer than that for the coherent motion though the difference was not
43
significant. This fact indicates that the motion type affected the response. The difference in the two types of motion is detected by the spatiotemporal integration of the dots’ local motions because there was no difference in the other visual information, such as the contrast, the dots’ size and density, temporal frequency, and local motion speed. Thus, the response must have occurred after the spatiotemporal integration of the stimuli as we discussed in our previous studies [11,12]. It is not plausible that the spatiotemporal integration occurs only in the detection of coherent motion and that incoherent motion is simply perceived as numerous local motions. First, the response latency for the incoherent motion was longer than that for the coherent motion. Because the neural process for the spatiotemporal integration is considered to occur after the local motion detection [19], the response latency for the coherent motion would be longer than that for the incoherent motion only if local motion detection is necessary to perceive incoherent motion. Second, the estimated source of the response to the incoherent motion was in the extrastriate area as was that of the coherent motion. The neurons in the extrastriate area including MT þ have larger classical receptive fields than those in the primary visual cortex [18]. This is why the major role of MT and the other extrastriate areas is considered to be that of integrating local motions to detect global motions [7,17]. Thus, the main response to the incoherent motion should originate from the primary visual cortex if there is no spatiotemporal integration process in the detection of incoherent motion. When the coherent motion was presented on the background of slower incoherent motion, the magnetic response to this stimulus, C(10) þ I(1.3), was not different from that to the coherent motion, C(10) þ S. This result was expected based on the idea that the global motion (coherent motion) is extracted from the background noise (incoherent motion) and is represented in the neural activity of the extrastriate area. If this is the case, we should have seen the same result in the stimuli of C(1.3) þ I(10). Even though the local speed of the coherent motion was slower than that of the background incoherent motion, the global motion energy for C(1.3) is larger than that for I(10) because the incoherent motion does not have a global direction. The response to C(1.3) þ I(10), however, was not different from the response to I(10) þ S. This could not have been because the slow coherent motion was not detected. All the subjects perceived the coherent motion in the stimulus of C(1.3) þ I(10) vividly because the 50% coherency was much higher than the detection threshold (usually a few percent) [1,5,14]. Furthermore, the results cannot be simply explained by the idea that the faster motions would be responsible for the magnetic response irrespective of the motion type. Although the local motion speed for I(10) was faster than that of C(1.3), the response is considered to occur after the spatiotemporal summation of the stimuli as discussed above. Again, the global motion energy for
44
K. Lam et al. / Neuroscience Letters 347 (2003) 41–44
C(1.3) is larger than that for I(10). Thus, the result indicates that coherent motion is not exclusively represented in the neural activity of the human extrastriate area and that the neural activity for incoherent motion also exists in that area. The results that the responses to the mixed motions were similar to those for the faster motions irrespective of the motion type show that the responses to the slower motions in the mixed motions were suppressed by the faster motions of the different types. Thus, the neural activity representing the coherent and incoherent motions may interact in the extrastriate area, though the two activities may occur in the different neural populations as suggested in our previous study [12].
Acknowledgements We thank Mr O. Nagata and Y. Takeshima for technical support.
References [1] C.L. Baker Jr., R.F. Hess, J. Zihl, Residual motion perception in a “motion-blind” patient, assessed with limited-lifetime random dot stimuli, J. Neurosci. 11 (1991) 454–461. [2] O.J. Braddick, J.M.D. O’Brien, J. Wattam Bell, J. Atkinson, T. Hartley, Brain areas sensitive to coherent visual motion, Perception 30 (2001) 61–72. [3] K.H. Britten, M.N. Shadlen, W.T. Newsome, J.A. Movshon, Responses of neurons in macaque MT to stochastic motion signals, Vis. Neurosci. 10 (1993) 1157–1169. [4] M. Bundo, Y. Kaneoke, S. Inao, J. Yoshida, A. Nakamura, R. Kakigi, Human visual motion areas determined individually by magnetoencephalography and 3D magnetic resonance imaging, Hum. Brain Mapp. 11 (2000) 33–45. [5] S. Celebrini, W.T. Newsome, Neuronal and psychophysical sensitivity to motion signals in extrastriate area MST of the macaque monkey, J. Neurosci. 14 (1994) 4109–4124.
[6] K. Cheng, H. Fujita, I. Kanno, S. Miura, K. Tanaka, Human cortical regions activated by wide-field visual motion: an H2(15)O PET study, J. Neurophysiol. 74 (1995) 413–427. [7] M.S. Graziano, R.A. Andersen, R.J. Snowden, Tuning of MST neurons to spiral motions, J. Neurosci. 14 (1994) 54–67. [8] D.J. Heeger, G.M. Boynton, J.B. Demb, E. Seidemann, W.T. Newsome, Motion opponency in visual cortex, J. Neurosci. 19 (1999) 7162– 7174. [9] K. Hol, S. Treue, Different populations of neurons contribute to the detection and discrimination of visual motion, Vis. Res. 41 (2001) 685 –689. [10] Y. Kaneoke, M. Bundou, S. Koyama, H. Suzuki, R. Kakigi, Human cortical area responding to stimuli in apparent motion, NeuroReport 8 (1997) 677 –682. [11] K. Lam, Y. Kaneoke, A. Gunji, H. Yamasaki, E. Matsumoto, T. Naito, R. Kakigi, Magnetic response of human extrastriate cortex in the detection of coherent and incoherent motion, Neuroscience 97 (2000) 1–10. [12] K. Maruyama, Y. Kaneoke, K. Watanabe, R. Kakigi, Human cortical responses to coherent and incoherent motion as measured by magnetoencephalography, Neurosci. Res. 44 (2002) 193–203. [13] D.J. McKeefry, J.D.G. Watson, R.S. Frackowiak, K. Fong, S. Zeki, The activity in human areas V1/V2, V3, and V5 during the perception of coherent and incoherent motion, Neuroimage 5 (1997) 1–12. [14] W.T. Newsome, K.H. Britten, J.A. Movshon, Neuronal correlates of a perceptual decision, Nature 341 (1989) 52–54. [15] W.T. Newsome, E.B. Pare, A selective impairment of motion perception following lesions of the middle temporal visual area (MT), J. Neurosci. 8 (1988) 2201– 2211. [16] G. Rees, K. Friston, C. Koch, A direct quantitative relationship between the functional properties of human and macaque V5, Nat. Neurosci. 3 (2000) 716 –723. [17] H. Sakata, H. Shibutani, Y. Ito, K. Tsurugai, Parietal cortical neurons responding to rotary movement of visual stimulus in space, Exp. Brain Res. 61 (1986) 658– 663. [18] D.C. Van Essen, J.H. Maunsell, J.L. Bixby, The middle temporal visual area in the macaque: myeloarchitecture, connections, functional properties and topographic organization, J. Comp. Neurol. 199 (1981) 293 –326. [19] D.W. Williams, R. Sekuler, Coherent global motion percepts from stochastic local motions, Vis. Res. 24 (1984) 55–62.