As TCM has been disseminated worldwide, different versions of acupuncture techniques have been developed, some of which do not follow the TCM principles for acupoint selection. For instance, 'trigger point acupuncture' involves stimulation of only Ashi points, single-point acupuncture uses only one acupoint for treatment and the eclectic approach involves needling anywhere within the region of the pain without any reference to Ashi or traditional acupoints [3]. Despite the various types of new acupuncture approaches, traditional point-specific techniques remain the most common form of acupuncture.

Minimal acupuncture utilizes needles inserted only 1-2 mm below the surface of the skin [10]. Needles are applied with no stimulation or light stimulation either mechanically by twirling the needles or electrically by applying a small pulsating current to the needles [10-14]. Despite the popular use of minimal acupuncture as a sham control, many clinical trials have suggested that this type of control has the potential to evoke analgesia through stimulation of cutaneous afferent nerves [15]. However, a recent study in mice reported that although superficial needle insertion caused analgesic effects, the effect was local and distinct from the acupuncture responses that rely on the afferent innervation of the skin [16]. In addition, the effects of electroacupuncture (EA) on blood pressure have been shown to result from stimulation of deeper nerves such as the median nerve located below Jianshi-Neiguan (P5-6) acupoints rather than from stimulation of superficial (i.e., cutaneous) branches such as the superficial radial nerve at Panli-Wenliu (LI6-LI7) acupoints [17,18]. In fact, investigation of reflex-induced pressor responses in cats showed that minimal or superficial acupuncture can serve as a valid control, because it does not alter the cardiovascular responses [19].

Needling at either inactive acupoints or non-acupoints are also common controls used in studies of acupuncture [20-24]. Non-acupoints are points located several millimeters or centimeters from verum acupoints, halfway between two parallel meridians or arbitrarily on the side of the trunk or the shoulder region away from most meridians [25]. Non-Ashi and non-acupoints are thought to have no therapeutic influence and thus may serve as inert placebo controls. Stimulation of inactive acupoints also is believed to be ineffective, so they can be used as sham controls. Inactive acupuncture points are chosen according to their anatomical locations, underlying neural pathways, corresponding Chinese meridians, proximity to verum acupoints and role in treating diseases [18,19,21,23]. With adequate blinding, the effects of stimulating either inactive acupoints or non-acupoints have been hypothesized to represent placebo responses.

The difference between sham and verum acupuncture remains unclear. Recent systematic reviews have suggested that stimulation of non-acupoints and inactive acupoints can elicit effects similar to those of stimulation of verum acupoints [26,27]. Thus, Langevin et al. [5] noted similar responses in sham and verum acupoint stimulation and attributed the nonspecific needling effects to either the summation of stimulation of multiple superficial sensory nerve endings or to the therapeutic effect related to practitioner-patient interactions. Despite these findings, some experts believe that it is illogical to conclude that stimulating all points on the body will elicit the same effect, because there is an uneven distribution of nerve endings throughout the body, suggesting differential sensory input at different points on the body [25]. In fact, many well-controlled studies do support the principle of point specificity. For example, recent studies in cardiovascular disease have shown that stimulation of verum acupoints elicits significantly greater responses than stimulation of both non-acupoints and inactive acupoints [18,19,23]. In addition, brain imaging studies have documented differential patterns of activation resulting from stimulation of inactive acupoints and non-acupoints [21,28]. Furthermore, experimental investigations have found that stimulation of different acupoints produces differential input to regions of the brain that regulate sympathetic outflow and cardiovascular function [18].

Similarly, other studies of acupuncture in acute pain have not been able to validate the concept of point specificity [34-37]. For example, Linde et al. [34] reported that measurement of headache intensity after stimulation of verum acupoints and non-acupoints in 208 subjects showed no difference in the effectiveness of pain reduction by either method. These studies were generally well controlled with adequate numbers of subjects, but the outcomes were measured by a visual analogue scale (VAS), which is highly variable between individuals [38]. Thus, any differences between placebo and nonspecific actions between sham and verum acupuncture may not have been detected.

Other hemodynamic studies also have compared the responses to stimulation of inactive acupoints with those of verum acupuncture stimulation. In this regard, Li et al. [23] examined systolic, diastolic and mean blood pressure, heart rate and the rate-pressure product (RPP) as an index of myocardial oxygen demand, in 17 subjects before and after exercise. The researchers measured the exercise-induced pressor responses immediately following EA stimulation of P5-6, LI4-L7 and Guangming-Xuanzhong (GB37-39) acupoints. They found that EA applied at acupoints P5-6 and LI4-L7 increased the maximum workload and reduced systolic and mean blood pressures as well as the RPP during exercise, while EA applied to GB37-39 did not alter these hemodynamic responses. They concluded that stimulation of inactive acupoints does not alter the sympathoexcitatory responses to exercise as seen with stimulation of verum acupoints.

Similar observations have been made in experimental studies of the responses to stimulation of LI6-7, P5-6, Yinxi-Shenmen (H6-7) and Zusanli-Shangjuxu (S36-37) acupoints on the pressor responses from gastric distension in rats, a model used for investigating visceral excitatory cardiovascular reflexes [19]. The control acupoints LI6-7 were chosen since they have not been used to treat cardiovascular disease and have little input to regions of the brain that regulate cardiovascular function [18]. The results showed that EA at LI6-7 did not modulate the reflex hemodynamic responses, while stimulation at P5-6, H6-7 and S36-37 significantly reduced the sympathoexcitatory responses. Thus, while multiple acupoints can be used to modulate cardiovascular function, the responses to EA are quite point-specific.

Stimulating different verum acupoints also has been shown to elicit differential patterns of brain activity. Zhang et al. [42] for example found that stimulating different sets of acupoints to treat different conditions elicited distinct patterns of activation in the brain. This study compared stimulation of acupoints S36 and Sanyinjiao (SP6), used to treat visceral disorders [43], to Yanglingquan (GB34) and Chengshan (B57), used for disorders of the muscles and tendons [44]. The fMRI data showed that stimulation of S36 and SP6 activated the orbital frontal cortex and deactivated the hippocampus and parietal BA7, regions associated with visceral disorders [45,46]. On the other hand, stimulation of GB34 and B57 increased activity in the dorsal thalamus and inhibited activity in the primary motor and premotor cortices. Interestingly, even though the four acupoints are located within the same spinal segment, they elicited different responses when stimulated. These results suggest that stimulation of different sets of acupoints leads to disease-specific neuronal responses, even when acupoints are located within the same spinal segment.