Table 1.
The examples of OM-SRC, OM-ORC, SM-SRC, and SM-ORC sentences.
Fig 1.
(A) The timeline of one sentence trial. Each trial starts with a fixation cross presented for 300 ms, followed by a 100 ms blank. Six frames are successively presented on the screen for 500 ms with 100 ms inter-stimulus interval. (B) The timeline of one run in the experiment. Each run contains six blocks, with two blocks of each type of sentences (OM-SRC and OM-ORC) and two blocks of the baseline condition (VO). The blocks in each run are in a pseudorandomized order. Note: VO = visual orientation; OM-SRC = object-modifying subject-extracted relative clause; OM-ORC = object-modifying object-extracted relative clause.
Fig 2.
The comparison of the comprehension performance in SM-SRC, SM-ORC, OM-SRC, and OM-ORC sentences.
The results show that the OM-RC sentences are more difficult to process than the SM-RC sentences. Among all sentences, OM-SRCs seems to be the most difficult one to comprehend, followed by OM-ORCs, SM-SRCs, and SM-ORCs.
Fig 3.
Volume rendering of SM-SRC, SM-ORC, OM-SRC and OM-ORC sentences.
The similar bilateral activation patterns are identified in response to the processing of the SRC and the ORC, regardless of the modifying position. The results show that (A) for the SM-RC sentences, only a left-lateralized frontotemporal neural network is involved; whereas (B) for the OM-RC sentences, besides the typical left-hemispheric brain areas, the right superior temporal gyrus (RSTG) is also recruited in sentence processing.
Table 2.
Locus and extent of peak activation in brain regions during sentence comprehension.
Fig 4.
Volume rendering of the contrast between the SRCs and the ORCs in different modifying positions.
The results indicate that (A) greater activation in the LIFG and the LSTG is significantly involved in the processing of the SM-SRC sentences compared to the SM-ORC sentences; (B) enhanced activation in the LIFG is significantly evoked by the OM-SRC sentences rather than the OM-ORC sentences.
Fig 5.
The simplified graph for the processing mechanism involved in the RC sentences with different levels of processing difficulty.
The results illustrate A) The activation of the LIFG and the LSTG, as well as the connectivity from the LIFG to the LSTG, is involved in the processing of the SM-SRCs; B) No evident functional communication between the LIFG and the LSTG is found in the reading of the SM-ORCs, although relatively less activation in the LIFG and the LSTG is still needed; C) The activation in the LIFG and also in the LSTG and its right homolog, as well as the effective connectivity from the LSTG to the LIFG, together facilitate the processing of the OM-SRCs; D) Compared with the OM-SRCs, the similar neural mechanism is involved in the reading of the OM-ORCs, although the activation of the LIFG is significantly less in the OM-ORCs compared to that in the OM-SRCs.